Add support for properly optimized Windows ARM64 builds with LLVM and MSVC (#7191)
* logging: add proper checks for clang to avoid errors and warnings with VA_ARGS * build: add CMake Presets and toolchian files for Windows ARM64 * matmul-int8: enable matmul-int8 with MSVC and fix Clang warnings * ci: add support for optimized Windows ARM64 builds with MSVC and LLVM * matmul-int8: fixed typos in q8_0_q8_0 matmuls Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> * matmul-int8: remove unnecessary casts in q8_0_q8_0 --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Max Krasnyansky
GitHub
parent
8f7080bf48
commit
13ad16af12
7 changed files with 138 additions and 58 deletions
61
.github/workflows/build.yml
vendored
61
.github/workflows/build.yml
vendored
|
|
@ -693,26 +693,28 @@ jobs:
|
|||
strategy:
|
||||
matrix:
|
||||
include:
|
||||
- build: 'rpc'
|
||||
- build: 'rpc-x64'
|
||||
defines: '-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_RPC=ON -DBUILD_SHARED_LIBS=ON'
|
||||
- build: 'noavx'
|
||||
- build: 'noavx-x64'
|
||||
defines: '-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_AVX=OFF -DLLAMA_AVX2=OFF -DLLAMA_FMA=OFF -DBUILD_SHARED_LIBS=ON'
|
||||
- build: 'avx2'
|
||||
- build: 'avx2-x64'
|
||||
defines: '-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DBUILD_SHARED_LIBS=ON'
|
||||
- build: 'avx'
|
||||
- build: 'avx-x64'
|
||||
defines: '-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_AVX2=OFF -DBUILD_SHARED_LIBS=ON'
|
||||
- build: 'avx512'
|
||||
- build: 'avx512-x64'
|
||||
defines: '-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_AVX512=ON -DBUILD_SHARED_LIBS=ON'
|
||||
- build: 'clblast'
|
||||
- build: 'clblast-x64'
|
||||
defines: '-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_CLBLAST=ON -DBUILD_SHARED_LIBS=ON -DCMAKE_PREFIX_PATH="$env:RUNNER_TEMP/clblast"'
|
||||
- build: 'openblas'
|
||||
- build: 'openblas-x64'
|
||||
defines: '-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_BLAS=ON -DBUILD_SHARED_LIBS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS -DBLAS_INCLUDE_DIRS="$env:RUNNER_TEMP/openblas/include" -DBLAS_LIBRARIES="$env:RUNNER_TEMP/openblas/lib/openblas.lib"'
|
||||
- build: 'kompute'
|
||||
- build: 'kompute-x64'
|
||||
defines: '-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_KOMPUTE=ON -DKOMPUTE_OPT_DISABLE_VULKAN_VERSION_CHECK=ON -DBUILD_SHARED_LIBS=ON'
|
||||
- build: 'vulkan'
|
||||
- build: 'vulkan-x64'
|
||||
defines: '-DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_VULKAN=ON -DBUILD_SHARED_LIBS=ON'
|
||||
- build: 'arm64'
|
||||
defines: '-A ARM64 -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DBUILD_SHARED_LIBS=ON'
|
||||
- build: 'llvm-arm64'
|
||||
defines: '-G Ninja -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-llvm.cmake -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DBUILD_SHARED_LIBS=ON'
|
||||
- build: 'msvc-arm64'
|
||||
defines: '-G Ninja -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-msvc.cmake -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DBUILD_SHARED_LIBS=ON'
|
||||
|
||||
steps:
|
||||
- name: Clone
|
||||
|
|
@ -723,13 +725,13 @@ jobs:
|
|||
|
||||
- name: Clone Kompute submodule
|
||||
id: clone_kompute
|
||||
if: ${{ matrix.build == 'kompute' }}
|
||||
if: ${{ matrix.build == 'kompute-x64' }}
|
||||
run: |
|
||||
git submodule update --init kompute
|
||||
|
||||
- name: Download OpenCL SDK
|
||||
id: get_opencl
|
||||
if: ${{ matrix.build == 'clblast' }}
|
||||
if: ${{ matrix.build == 'clblast-x64' }}
|
||||
run: |
|
||||
curl.exe -o $env:RUNNER_TEMP/opencl.zip -L "https://github.com/KhronosGroup/OpenCL-SDK/releases/download/v${env:OPENCL_VERSION}/OpenCL-SDK-v${env:OPENCL_VERSION}-Win-x64.zip"
|
||||
mkdir $env:RUNNER_TEMP/opencl
|
||||
|
|
@ -737,7 +739,7 @@ jobs:
|
|||
|
||||
- name: Download CLBlast
|
||||
id: get_clblast
|
||||
if: ${{ matrix.build == 'clblast' }}
|
||||
if: ${{ matrix.build == 'clblast-x64' }}
|
||||
run: |
|
||||
curl.exe -o $env:RUNNER_TEMP/clblast.7z -L "https://github.com/CNugteren/CLBlast/releases/download/${env:CLBLAST_VERSION}/CLBlast-${env:CLBLAST_VERSION}-windows-x64.7z"
|
||||
curl.exe -o $env:RUNNER_TEMP/CLBlast.LICENSE.txt -L "https://github.com/CNugteren/CLBlast/raw/${env:CLBLAST_VERSION}/LICENSE"
|
||||
|
|
@ -750,7 +752,7 @@ jobs:
|
|||
|
||||
- name: Download OpenBLAS
|
||||
id: get_openblas
|
||||
if: ${{ matrix.build == 'openblas' }}
|
||||
if: ${{ matrix.build == 'openblas-x64' }}
|
||||
run: |
|
||||
curl.exe -o $env:RUNNER_TEMP/openblas.zip -L "https://github.com/xianyi/OpenBLAS/releases/download/v${env:OPENBLAS_VERSION}/OpenBLAS-${env:OPENBLAS_VERSION}-x64.zip"
|
||||
curl.exe -o $env:RUNNER_TEMP/OpenBLAS.LICENSE.txt -L "https://github.com/xianyi/OpenBLAS/raw/v${env:OPENBLAS_VERSION}/LICENSE"
|
||||
|
|
@ -763,38 +765,41 @@ jobs:
|
|||
|
||||
- name: Install Vulkan SDK
|
||||
id: get_vulkan
|
||||
if: ${{ matrix.build == 'kompute' || matrix.build == 'vulkan' }}
|
||||
if: ${{ matrix.build == 'kompute-x64' || matrix.build == 'vulkan-x64' }}
|
||||
run: |
|
||||
curl.exe -o $env:RUNNER_TEMP/VulkanSDK-Installer.exe -L "https://sdk.lunarg.com/sdk/download/${env:VULKAN_VERSION}/windows/VulkanSDK-${env:VULKAN_VERSION}-Installer.exe"
|
||||
& "$env:RUNNER_TEMP\VulkanSDK-Installer.exe" --accept-licenses --default-answer --confirm-command install
|
||||
Add-Content $env:GITHUB_ENV "VULKAN_SDK=C:\VulkanSDK\${env:VULKAN_VERSION}"
|
||||
Add-Content $env:GITHUB_PATH "C:\VulkanSDK\${env:VULKAN_VERSION}\bin"
|
||||
|
||||
- name: Install Ninja
|
||||
id: install_ninja
|
||||
run: |
|
||||
choco install ninja
|
||||
|
||||
- name: Build
|
||||
id: cmake_build
|
||||
run: |
|
||||
mkdir build
|
||||
cd build
|
||||
cmake .. ${{ matrix.defines }}
|
||||
cmake --build . --config Release -j ${env:NUMBER_OF_PROCESSORS}
|
||||
cmake -S . -B build ${{ matrix.defines }}
|
||||
cmake --build build --config Release -j ${env:NUMBER_OF_PROCESSORS}
|
||||
|
||||
- name: Add clblast.dll
|
||||
id: add_clblast_dll
|
||||
if: ${{ matrix.build == 'clblast' }}
|
||||
if: ${{ matrix.build == 'clblast-x64' }}
|
||||
run: |
|
||||
cp $env:RUNNER_TEMP/clblast/lib/clblast.dll ./build/bin/Release
|
||||
cp $env:RUNNER_TEMP/CLBlast.LICENSE.txt ./build/bin/Release/CLBlast-${env:CLBLAST_VERSION}.txt
|
||||
|
||||
- name: Add libopenblas.dll
|
||||
id: add_libopenblas_dll
|
||||
if: ${{ matrix.build == 'openblas' }}
|
||||
if: ${{ matrix.build == 'openblas-x64' }}
|
||||
run: |
|
||||
cp $env:RUNNER_TEMP/openblas/bin/libopenblas.dll ./build/bin/Release/openblas.dll
|
||||
cp $env:RUNNER_TEMP/OpenBLAS.LICENSE.txt ./build/bin/Release/OpenBLAS-${env:OPENBLAS_VERSION}.txt
|
||||
|
||||
- name: Check AVX512F support
|
||||
id: check_avx512f
|
||||
if: ${{ matrix.build == 'avx512' }}
|
||||
if: ${{ matrix.build == 'avx512-x64' }}
|
||||
continue-on-error: true
|
||||
run: |
|
||||
cd build
|
||||
|
|
@ -808,14 +813,14 @@ jobs:
|
|||
- name: Test
|
||||
id: cmake_test
|
||||
# not all machines have native AVX-512
|
||||
if: ${{ matrix.build != 'arm64' && matrix.build != 'clblast' && matrix.build != 'kompute' && matrix.build != 'vulkan' && (matrix.build != 'avx512' || env.HAS_AVX512F == '1') }}
|
||||
if: ${{ matrix.build != 'msvc-arm64' && matrix.build != 'llvm-arm64' && matrix.build != 'clblast-x64' && matrix.build != 'kompute-x64' && matrix.build != 'vulkan-x64' && (matrix.build != 'avx512-x64' || env.HAS_AVX512F == '1') }}
|
||||
run: |
|
||||
cd build
|
||||
ctest -L main -C Release --verbose --timeout 900
|
||||
|
||||
- name: Test (Intel SDE)
|
||||
id: cmake_test_sde
|
||||
if: ${{ matrix.build == 'avx512' && env.HAS_AVX512F == '0' }} # use Intel SDE for AVX-512 emulation
|
||||
if: ${{ matrix.build == 'avx512-x64' && env.HAS_AVX512F == '0' }} # use Intel SDE for AVX-512 emulation
|
||||
run: |
|
||||
curl.exe -o $env:RUNNER_TEMP/sde.tar.xz -L "https://downloadmirror.intel.com/813591/sde-external-${env:SDE_VERSION}-win.tar.xz"
|
||||
# for some weird reason windows tar doesn't like sde tar.xz
|
||||
|
|
@ -843,14 +848,14 @@ jobs:
|
|||
if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
|
||||
run: |
|
||||
Copy-Item LICENSE .\build\bin\Release\llama.cpp.txt
|
||||
7z a llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}-x64.zip .\build\bin\Release\*
|
||||
7z a llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}.zip .\build\bin\Release\*
|
||||
|
||||
- name: Upload artifacts
|
||||
if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
|
||||
uses: actions/upload-artifact@v4
|
||||
with:
|
||||
path: llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}-x64.zip
|
||||
name: llama-bin-win-${{ matrix.build }}-x64.zip
|
||||
path: llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}.zip
|
||||
name: llama-bin-win-${{ matrix.build }}.zip
|
||||
|
||||
windows-latest-cmake-cuda:
|
||||
runs-on: windows-latest
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue