From ff4cb6ef4c6113a6f8de439298a89dba4c6540cc Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 13:28:37 +0200
Subject: [PATCH 01/15] release : pack /lib and /include in the packages

---
 .github/workflows/build.yml | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index fe3b2cdfa..07e54751b 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -88,7 +88,7 @@ jobs:
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
-          zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.zip ./build/bin/*
+          zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.zip ./build/bin/* ./build/lib/* ./build/include/*
 
       - name: Upload artifacts
         if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
@@ -151,7 +151,7 @@ jobs:
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
-          zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip ./build/bin/*
+          zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip ./build/bin/* ./build/lib/* ./build/include/*
 
       - name: Upload artifacts
         if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
@@ -220,7 +220,7 @@ jobs:
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
-          zip -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-x64.zip ./build/bin/*
+          zip -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-x64.zip ./build/bin/* ./build/lib/* ./build/include/*
 
       - name: Upload artifacts
         if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
@@ -1026,7 +1026,7 @@ jobs:
           cp "${{ env.ONEAPI_ROOT }}/tbb/latest/bin/tbb12.dll" ./build/bin
 
           echo "cp oneAPI running time dll files to ./build/bin done"
-          7z a llama-${{ steps.tag.outputs.name }}-bin-win-sycl-x64.zip ./build/bin/*
+          7z a llama-${{ steps.tag.outputs.name }}-bin-win-sycl-x64.zip ./build/bin/* ./build/lib/* ./build/include/*
 
       - name: Upload the release package
         if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}

From 3a35bfe1f798061bb9b7a5679e48ae4e86261807 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 14:40:48 +0200
Subject: [PATCH 02/15] cmake : put libs in /bin

---
 .github/workflows/build.yml | 86 +++++++++++++++++++++++++++++--------
 CMakeLists.txt              |  1 +
 2 files changed, 68 insertions(+), 19 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 07e54751b..00d17e2f8 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -88,7 +88,7 @@ jobs:
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
-          zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.zip ./build/bin/* ./build/lib/* ./build/include/*
+          zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.zip ./build/bin/*
 
       - name: Upload artifacts
         if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
@@ -151,7 +151,7 @@ jobs:
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
-          zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip ./build/bin/* ./build/lib/* ./build/include/*
+          zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip ./build/bin/*
 
       - name: Upload artifacts
         if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
@@ -181,7 +181,10 @@ jobs:
         run: |
           mkdir build
           cd build
-          cmake .. -DLLAMA_FATAL_WARNINGS=ON -DLLAMA_CURL=ON -DGGML_RPC=ON
+          cmake .. \
+            -DLLAMA_FATAL_WARNINGS=ON \
+            -DLLAMA_CURL=ON \
+            -DGGML_RPC=ON
           cmake --build . --config Release -j $(nproc)
 
       - name: Test
@@ -220,7 +223,7 @@ jobs:
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
-          zip -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-x64.zip ./build/bin/* ./build/lib/* ./build/include/*
+          zip -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-x64.zip ./build/bin/*
 
       - name: Upload artifacts
         if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
@@ -256,7 +259,10 @@ jobs:
         run: |
           mkdir build
           cd build
-          cmake .. -DLLAMA_FATAL_WARNINGS=ON -DLLAMA_SANITIZE_${{ matrix.sanitizer }}=ON -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
+          cmake .. \
+            -DLLAMA_FATAL_WARNINGS=ON \
+            -DLLAMA_SANITIZE_${{ matrix.sanitizer }}=ON \
+            -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
           cmake --build . --config ${{ matrix.build_type }} -j $(nproc)
 
       - name: Build (no OpenMP)
@@ -265,7 +271,11 @@ jobs:
         run: |
           mkdir build
           cd build
-          cmake .. -DLLAMA_FATAL_WARNINGS=ON -DLLAMA_SANITIZE_${{ matrix.sanitizer }}=ON -DCMAKE_BUILD_TYPE=${{ matrix.build_type }} -DGGML_OPENMP=OFF
+          cmake .. \
+            -DLLAMA_FATAL_WARNINGS=ON \
+            -DLLAMA_SANITIZE_${{ matrix.sanitizer }}=ON \
+            -DCMAKE_BUILD_TYPE=${{ matrix.build_type }} \
+            -DGGML_OPENMP=OFF
           cmake --build . --config ${{ matrix.build_type }} -j $(nproc)
 
       - name: Test
@@ -295,7 +305,8 @@ jobs:
         run: |
           mkdir build
           cd build
-          cmake -DGGML_RPC=ON ..
+          cmake .. \
+            -DGGML_RPC=ON
           cmake --build . --config Release -j $(nproc)
 
       - name: Test
@@ -325,7 +336,8 @@ jobs:
         run: |
           mkdir build
           cd build
-          cmake -DGGML_VULKAN=ON ..
+          cmake .. \
+            -DGGML_VULKAN=ON
           cmake --build . --config Release -j $(nproc)
 
       - name: Test
@@ -352,13 +364,18 @@ jobs:
       - name: Build with native CMake HIP support
         id: cmake_build
         run: |
-          cmake -B build -S . -DCMAKE_HIP_COMPILER="$(hipconfig -l)/clang" -DGGML_HIP=ON
+          cmake -B build -S . \
+            -DCMAKE_HIP_COMPILER="$(hipconfig -l)/clang" \
+            -DGGML_HIP=ON
           cmake --build build --config Release -j $(nproc)
 
       - name: Build with legacy HIP support
         id: cmake_build_legacy_hip
         run: |
-          cmake -B build2 -S . -DCMAKE_C_COMPILER=hipcc -DCMAKE_CXX_COMPILER=hipcc -DGGML_HIP=ON
+          cmake -B build2 -S . \
+            -DCMAKE_C_COMPILER=hipcc \
+            -DCMAKE_CXX_COMPILER=hipcc \
+            -DGGML_HIP=ON
           cmake --build build2 --config Release -j $(nproc)
 
   ubuntu-22-cmake-musa:
@@ -379,7 +396,8 @@ jobs:
       - name: Build with native CMake MUSA support
         id: cmake_build
         run: |
-          cmake -B build -S . -DGGML_MUSA=ON
+          cmake -B build -S . \
+            -DGGML_MUSA=ON
           cmake --build build --config Release -j $(nproc)
 
   ubuntu-22-cmake-sycl:
@@ -420,7 +438,10 @@ jobs:
           source /opt/intel/oneapi/setvars.sh
           mkdir build
           cd build
-          cmake -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx ..
+          cmake .. \
+            -DGGML_SYCL=ON \
+            -DCMAKE_C_COMPILER=icx \
+            -DCMAKE_CXX_COMPILER=icpx
           cmake --build . --config Release -j $(nproc)
 
   ubuntu-22-cmake-sycl-fp16:
@@ -461,7 +482,11 @@ jobs:
           source /opt/intel/oneapi/setvars.sh
           mkdir build
           cd build
-          cmake -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_SYCL_F16=ON ..
+          cmake .. \
+            -DGGML_SYCL=ON \
+            -DCMAKE_C_COMPILER=icx \
+            -DCMAKE_CXX_COMPILER=icpx \
+            -DGGML_SYCL_F16=ON
           cmake --build . --config Release -j $(nproc)
 
   # TODO: build with GGML_METAL=OFF because test-backend-ops fail on "Apple Paravirtual device" and I don't know
@@ -488,7 +513,9 @@ jobs:
           sysctl -a
           mkdir build
           cd build
-          cmake -DLLAMA_FATAL_WARNINGS=ON -DGGML_METAL=OFF ..
+          cmake .. \
+            -DLLAMA_FATAL_WARNINGS=ON \
+            -DGGML_METAL=OFF
           cmake --build . --config Release -j $(sysctl -n hw.logicalcpu)
 
       - name: Test
@@ -827,7 +854,13 @@ jobs:
 
         - name: Build with CMake
           run: |
-            cmake -S . -B build -G Ninja -DCMAKE_BUILD_TYPE=Release -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=89-real -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined -DLLAMA_FATAL_WARNINGS=ON
+            cmake -S . -B build -G Ninja \
+              -DCMAKE_BUILD_TYPE=Release \
+              -DCMAKE_CUDA_ARCHITECTURES=89-real \
+              -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined \
+              -DLLAMA_FATAL_WARNINGS=ON \
+              -DGGML_NATIVE=OFF \
+              -DGGML_CUDA=ON
             cmake --build build
 
   windows-2019-cmake-cuda:
@@ -916,7 +949,11 @@ jobs:
         shell: cmd
         run: |
           call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat"
-          cmake -S . -B build -G "Ninja Multi-Config" -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DGGML_RPC=ON
+          cmake -S . -B build -G "Ninja Multi-Config" \
+            -DLLAMA_BUILD_SERVER=ON \
+            -DGGML_NATIVE=OFF \
+            -DGGML_CUDA=ON \
+            -DGGML_RPC=ON
           set /A NINJA_JOBS=%NUMBER_OF_PROCESSORS%-1
           cmake --build build --config Release -j %NINJA_JOBS% -t ggml
           cmake --build build --config Release
@@ -1026,7 +1063,7 @@ jobs:
           cp "${{ env.ONEAPI_ROOT }}/tbb/latest/bin/tbb12.dll" ./build/bin
 
           echo "cp oneAPI running time dll files to ./build/bin done"
-          7z a llama-${{ steps.tag.outputs.name }}-bin-win-sycl-x64.zip ./build/bin/* ./build/lib/* ./build/include/*
+          7z a llama-${{ steps.tag.outputs.name }}-bin-win-sycl-x64.zip ./build/bin/*
 
       - name: Upload the release package
         if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
@@ -1069,7 +1106,12 @@ jobs:
         run: |
           $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
           $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}"
-          cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIP=ON -DCMAKE_BUILD_TYPE=Release -DGGML_RPC=ON
+          cmake -G "Unix Makefiles" -B build -S . \
+            -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" \
+            -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" \
+            -DCMAKE_BUILD_TYPE=Release \
+            -DGGML_HIP=ON \
+            -DGGML_RPC=ON
           cmake --build build -j ${env:NUMBER_OF_PROCESSORS}
 
   windows-latest-cmake-hip-release:
@@ -1107,7 +1149,13 @@ jobs:
         run: |
           $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
           $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}"
-          cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIP=ON -DCMAKE_BUILD_TYPE=Release -DAMDGPU_TARGETS=${{ matrix.gpu_target }} -DGGML_RPC=ON
+          cmake -G "Unix Makefiles" -B build -S . \
+            -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" \
+            -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" \
+            -DCMAKE_BUILD_TYPE=Release \
+            -DAMDGPU_TARGETS=${{ matrix.gpu_target }} \
+            -DGGML_HIP=ON \
+            -DGGML_RPC=ON
           cmake --build build -j ${env:NUMBER_OF_PROCESSORS}
           md "build\bin\rocblas\library\"
           cp "${env:HIP_PATH}\bin\hipblas.dll" "build\bin\"
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 42caed486..7e41a44d2 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -16,6 +16,7 @@ endif()
 list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/cmake/")
 
 set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)
+set(CMAKE_LIBRARY_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)
 
 if (CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR)
     set(LLAMA_STANDALONE ON)

From 4decf2c4dfc5cdf5d96ea44c03c8f9801ab41262 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 14:54:24 +0200
Subject: [PATCH 03/15] TMP : push artifacts

---
 .github/workflows/build.yml | 14 --------------
 1 file changed, 14 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 00d17e2f8..d28fcc067 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -84,14 +84,12 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.zip ./build/bin/*
 
       - 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-macos-arm64.zip
@@ -147,14 +145,12 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip ./build/bin/*
 
       - 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-macos-x64.zip
@@ -219,14 +215,12 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-x64.zip ./build/bin/*
 
       - 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-ubuntu-x64.zip
@@ -823,14 +817,12 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           Copy-Item LICENSE .\build\bin\Release\llama.cpp.txt
           Copy-Item .\examples\run\linenoise.cpp\LICENSE .\build\bin\Release\linenoise.cpp.txt
           7z a llama-${{ 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 }}.zip
@@ -973,19 +965,16 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           7z a llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}-cu${{ matrix.cuda }}-x64.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-${{ matrix.build }}-cu${{ matrix.cuda }}-x64.zip
           name: llama-bin-win-cu${{ matrix.cuda }}-x64.zip
 
       - name: Copy and pack Cuda runtime
-        if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
         run: |
           echo "Cuda install location: ${{ env.CUDA_PATH }}"
           $dst='.\build\bin\cudart\'
@@ -994,7 +983,6 @@ jobs:
           7z a cudart-llama-bin-win-cu${{ matrix.cuda }}-x64.zip $dst\*
 
       - name: Upload Cuda runtime
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         uses: actions/upload-artifact@v4
         with:
           path: cudart-llama-bin-win-cu${{ matrix.cuda }}-x64.zip
@@ -1041,7 +1029,6 @@ jobs:
 
       - name: Build the release package
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           echo "cp oneAPI running time dll files in ${{ env.ONEAPI_ROOT }} to ./build/bin"
 
@@ -1066,7 +1053,6 @@ jobs:
           7z a llama-${{ steps.tag.outputs.name }}-bin-win-sycl-x64.zip ./build/bin/*
 
       - name: Upload the release package
-        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-sycl-x64.zip

From 564353c9a31c0149b1a95a50d691ff59960310ec Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 15:22:36 +0200
Subject: [PATCH 04/15] Revert "TMP : push artifacts"

This reverts commit 4decf2c4dfc5cdf5d96ea44c03c8f9801ab41262.
---
 .github/workflows/build.yml | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index d28fcc067..00d17e2f8 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -84,12 +84,14 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.zip ./build/bin/*
 
       - 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-macos-arm64.zip
@@ -145,12 +147,14 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip ./build/bin/*
 
       - 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-macos-x64.zip
@@ -215,12 +219,14 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-x64.zip ./build/bin/*
 
       - 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-ubuntu-x64.zip
@@ -817,12 +823,14 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           Copy-Item LICENSE .\build\bin\Release\llama.cpp.txt
           Copy-Item .\examples\run\linenoise.cpp\LICENSE .\build\bin\Release\linenoise.cpp.txt
           7z a llama-${{ 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 }}.zip
@@ -965,16 +973,19 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           7z a llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}-cu${{ matrix.cuda }}-x64.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-${{ matrix.build }}-cu${{ matrix.cuda }}-x64.zip
           name: llama-bin-win-cu${{ matrix.cuda }}-x64.zip
 
       - name: Copy and pack Cuda runtime
+        if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
         run: |
           echo "Cuda install location: ${{ env.CUDA_PATH }}"
           $dst='.\build\bin\cudart\'
@@ -983,6 +994,7 @@ jobs:
           7z a cudart-llama-bin-win-cu${{ matrix.cuda }}-x64.zip $dst\*
 
       - name: Upload Cuda runtime
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         uses: actions/upload-artifact@v4
         with:
           path: cudart-llama-bin-win-cu${{ matrix.cuda }}-x64.zip
@@ -1029,6 +1041,7 @@ jobs:
 
       - name: Build the release package
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           echo "cp oneAPI running time dll files in ${{ env.ONEAPI_ROOT }} to ./build/bin"
 
@@ -1053,6 +1066,7 @@ jobs:
           7z a llama-${{ steps.tag.outputs.name }}-bin-win-sycl-x64.zip ./build/bin/*
 
       - name: Upload the release package
+        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-sycl-x64.zip

From 50455ded315d063b5be6a4811c9b398ef54a7d47 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 15:23:22 +0200
Subject: [PATCH 05/15] ci : fix HIP cmake compiler options to be on first line

---
 .github/workflows/build.yml | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 00d17e2f8..953bf8063 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -1106,9 +1106,7 @@ jobs:
         run: |
           $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
           $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}"
-          cmake -G "Unix Makefiles" -B build -S . \
-            -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" \
-            -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" \
+          cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" \
             -DCMAKE_BUILD_TYPE=Release \
             -DGGML_HIP=ON \
             -DGGML_RPC=ON
@@ -1149,9 +1147,7 @@ jobs:
         run: |
           $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
           $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}"
-          cmake -G "Unix Makefiles" -B build -S . \
-            -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" \
-            -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" \
+          cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" \
             -DCMAKE_BUILD_TYPE=Release \
             -DAMDGPU_TARGETS=${{ matrix.gpu_target }} \
             -DGGML_HIP=ON \

From 194358e3b7b42206faab26edeb2fa98040de3de8 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 15:41:52 +0200
Subject: [PATCH 06/15] ci : restore the original HIP commands

---
 .github/workflows/build.yml | 11 ++---------
 1 file changed, 2 insertions(+), 9 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 953bf8063..172c01d33 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -1106,10 +1106,7 @@ jobs:
         run: |
           $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
           $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}"
-          cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" \
-            -DCMAKE_BUILD_TYPE=Release \
-            -DGGML_HIP=ON \
-            -DGGML_RPC=ON
+          cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIP=ON -DCMAKE_BUILD_TYPE=Release -DGGML_RPC=ON
           cmake --build build -j ${env:NUMBER_OF_PROCESSORS}
 
   windows-latest-cmake-hip-release:
@@ -1147,11 +1144,7 @@ jobs:
         run: |
           $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
           $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}"
-          cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" \
-            -DCMAKE_BUILD_TYPE=Release \
-            -DAMDGPU_TARGETS=${{ matrix.gpu_target }} \
-            -DGGML_HIP=ON \
-            -DGGML_RPC=ON
+          cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIP=ON -DCMAKE_BUILD_TYPE=Release -DAMDGPU_TARGETS=${{ matrix.gpu_target }} -DGGML_RPC=ON
           cmake --build build -j ${env:NUMBER_OF_PROCESSORS}
           md "build\bin\rocblas\library\"
           cp "${env:HIP_PATH}\bin\hipblas.dll" "build\bin\"

From 56e26a7f30c3f196577b2c2058a53cf18c2d4020 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 15:58:48 +0200
Subject: [PATCH 07/15] ci : change ubuntu build from latest to 20.04

---
 .github/workflows/build.yml | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 172c01d33..0c358b6cc 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -160,8 +160,8 @@ jobs:
           path: llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip
           name: llama-bin-macos-x64.zip
 
-  ubuntu-latest-cmake:
-    runs-on: ubuntu-latest
+  ubuntu-cmake:
+    runs-on: ubuntu-20.04
 
     steps:
       - name: Clone
@@ -1238,7 +1238,7 @@ jobs:
     runs-on: ubuntu-latest
 
     needs:
-      - ubuntu-latest-cmake
+      - ubuntu-cmake
       - macOS-latest-cmake
       - windows-latest-cmake
       - windows-2019-cmake-cuda

From f9f65f016299bd9713855850cc3fe3f2f7465f72 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 16:01:32 +0200
Subject: [PATCH 08/15] ci : try to fix macos build rpaths

---
 .github/workflows/build.yml | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 0c358b6cc..45c203e2b 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -56,6 +56,7 @@ jobs:
           mkdir build
           cd build
           cmake .. \
+            -DCMAKE_BUILD_RPATH_USE_ORIGIN=ON \
             -DLLAMA_FATAL_WARNINGS=ON \
             -DLLAMA_CURL=ON \
             -DGGML_METAL_USE_BF16=ON \
@@ -120,6 +121,7 @@ jobs:
           # Metal is disabled due to intermittent failures with Github runners not having a GPU:
           # https://github.com/ggerganov/llama.cpp/actions/runs/8635935781/job/23674807267#step:5:2313
           cmake -B build \
+            -DCMAKE_BUILD_RPATH_USE_ORIGIN=ON \
             -DLLAMA_FATAL_WARNINGS=ON \
             -DLLAMA_CURL=ON \
             -DGGML_METAL=OFF \

From 8b2ed1e432b0720fe1b6a06a03c492fd21df2cc7 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 16:01:52 +0200
Subject: [PATCH 09/15] ci : remove obsolete MacOS build

---
 .github/workflows/build.yml | 36 ------------------------------------
 1 file changed, 36 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 45c203e2b..591b8d702 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -491,41 +491,6 @@ jobs:
             -DGGML_SYCL_F16=ON
           cmake --build . --config Release -j $(nproc)
 
-  # TODO: build with GGML_METAL=OFF because test-backend-ops fail on "Apple Paravirtual device" and I don't know
-  #       how to debug it.
-  #       ref: https://github.com/ggerganov/llama.cpp/actions/runs/7132125951/job/19422043567?pr=4359#step:5:6584
-  #       would be great if we fix these
-  macOS-latest-cmake:
-    runs-on: macos-latest
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: Dependencies
-        id: depends
-        continue-on-error: true
-        run: |
-          brew update
-
-      - name: Build
-        id: cmake_build
-        run: |
-          sysctl -a
-          mkdir build
-          cd build
-          cmake .. \
-            -DLLAMA_FATAL_WARNINGS=ON \
-            -DGGML_METAL=OFF
-          cmake --build . --config Release -j $(sysctl -n hw.logicalcpu)
-
-      - name: Test
-        id: cmake_test
-        run: |
-          cd build
-          ctest -L main --verbose --timeout 900
-
   macOS-latest-cmake-ios:
     runs-on: macos-latest
 
@@ -1241,7 +1206,6 @@ jobs:
 
     needs:
       - ubuntu-cmake
-      - macOS-latest-cmake
       - windows-latest-cmake
       - windows-2019-cmake-cuda
       - windows-latest-cmake-hip-release

From 537b09e70ffc604c414ee78acf3acb4c940ec597 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 14:54:24 +0200
Subject: [PATCH 10/15] TMP : push artifacts

---
 .github/workflows/build.yml | 14 --------------
 1 file changed, 14 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 591b8d702..b84b32f24 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -85,14 +85,12 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.zip ./build/bin/*
 
       - 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-macos-arm64.zip
@@ -149,14 +147,12 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip ./build/bin/*
 
       - 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-macos-x64.zip
@@ -221,14 +217,12 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-x64.zip ./build/bin/*
 
       - 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-ubuntu-x64.zip
@@ -790,14 +784,12 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           Copy-Item LICENSE .\build\bin\Release\llama.cpp.txt
           Copy-Item .\examples\run\linenoise.cpp\LICENSE .\build\bin\Release\linenoise.cpp.txt
           7z a llama-${{ 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 }}.zip
@@ -940,19 +932,16 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           7z a llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}-cu${{ matrix.cuda }}-x64.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-${{ matrix.build }}-cu${{ matrix.cuda }}-x64.zip
           name: llama-bin-win-cu${{ matrix.cuda }}-x64.zip
 
       - name: Copy and pack Cuda runtime
-        if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
         run: |
           echo "Cuda install location: ${{ env.CUDA_PATH }}"
           $dst='.\build\bin\cudart\'
@@ -961,7 +950,6 @@ jobs:
           7z a cudart-llama-bin-win-cu${{ matrix.cuda }}-x64.zip $dst\*
 
       - name: Upload Cuda runtime
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         uses: actions/upload-artifact@v4
         with:
           path: cudart-llama-bin-win-cu${{ matrix.cuda }}-x64.zip
@@ -1008,7 +996,6 @@ jobs:
 
       - name: Build the release package
         id: pack_artifacts
-        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           echo "cp oneAPI running time dll files in ${{ env.ONEAPI_ROOT }} to ./build/bin"
 
@@ -1033,7 +1020,6 @@ jobs:
           7z a llama-${{ steps.tag.outputs.name }}-bin-win-sycl-x64.zip ./build/bin/*
 
       - name: Upload the release package
-        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-sycl-x64.zip

From dae44bf21a0e545b306fa9f67a346042cfc37f6d Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 16:28:32 +0200
Subject: [PATCH 11/15] ci : change back to ubuntu latest

---
 .github/workflows/build.yml | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index b84b32f24..bdf7cf4f3 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -158,8 +158,8 @@ jobs:
           path: llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip
           name: llama-bin-macos-x64.zip
 
-  ubuntu-cmake:
-    runs-on: ubuntu-20.04
+  ubuntu--latest-cmake:
+    runs-on: ubuntu--latest
 
     steps:
       - name: Clone
@@ -1191,7 +1191,7 @@ jobs:
     runs-on: ubuntu-latest
 
     needs:
-      - ubuntu-cmake
+      - ubuntu-latest-cmake
       - windows-latest-cmake
       - windows-2019-cmake-cuda
       - windows-latest-cmake-hip-release

From 39d0621872655e4e87866f0e321c6e6de627cf77 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 16:31:12 +0200
Subject: [PATCH 12/15] ci : macos set build rpath to "@loader_path"

---
 .github/workflows/build.yml | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index bdf7cf4f3..e15fe4080 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -56,7 +56,7 @@ jobs:
           mkdir build
           cd build
           cmake .. \
-            -DCMAKE_BUILD_RPATH_USE_ORIGIN=ON \
+            -DCMAKE_BUILD_RPATH="@loader_path" \
             -DLLAMA_FATAL_WARNINGS=ON \
             -DLLAMA_CURL=ON \
             -DGGML_METAL_USE_BF16=ON \
@@ -119,7 +119,7 @@ jobs:
           # Metal is disabled due to intermittent failures with Github runners not having a GPU:
           # https://github.com/ggerganov/llama.cpp/actions/runs/8635935781/job/23674807267#step:5:2313
           cmake -B build \
-            -DCMAKE_BUILD_RPATH_USE_ORIGIN=ON \
+            -DCMAKE_BUILD_RPATH="@loader_path" \
             -DLLAMA_FATAL_WARNINGS=ON \
             -DLLAMA_CURL=ON \
             -DGGML_METAL=OFF \

From 872fd18420a4bef96fdaa7cbd1b5d6f27d45804c Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 16:46:03 +0200
Subject: [PATCH 13/15] ci : fix typo

---
 .github/workflows/build.yml | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index e15fe4080..492991741 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -158,8 +158,8 @@ jobs:
           path: llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip
           name: llama-bin-macos-x64.zip
 
-  ubuntu--latest-cmake:
-    runs-on: ubuntu--latest
+  ubuntu-latest-cmake:
+    runs-on: ubuntu-latest
 
     steps:
       - name: Clone

From 5740ec7a661642aaeb07362efaabd397404dd5f5 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 17:05:44 +0200
Subject: [PATCH 14/15] ci : change ubuntu package to 22.04

---
 .github/workflows/build.yml | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 492991741..fd9dd783f 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -158,8 +158,8 @@ jobs:
           path: llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip
           name: llama-bin-macos-x64.zip
 
-  ubuntu-latest-cmake:
-    runs-on: ubuntu-latest
+  ubuntu-cpu-cmake:
+    runs-on: ubuntu-22.04
 
     steps:
       - name: Clone
@@ -1191,7 +1191,7 @@ jobs:
     runs-on: ubuntu-latest
 
     needs:
-      - ubuntu-latest-cmake
+      - ubuntu-cpu-cmake
       - windows-latest-cmake
       - windows-2019-cmake-cuda
       - windows-latest-cmake-hip-release

From 969b264657f12e42e1e3a2988bd790c8c2ee811a Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Fri, 24 Jan 2025 17:58:09 +0200
Subject: [PATCH 15/15] Revert "TMP : push artifacts"

This reverts commit 537b09e70ffc604c414ee78acf3acb4c940ec597.
---
 .github/workflows/build.yml | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index fd9dd783f..fda726955 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -85,12 +85,14 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.zip ./build/bin/*
 
       - 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-macos-arm64.zip
@@ -147,12 +149,14 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip ./build/bin/*
 
       - 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-macos-x64.zip
@@ -217,12 +221,14 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           cp LICENSE ./build/bin/
           cp examples/run/linenoise.cpp/LICENSE ./build/bin/LICENSE.linenoise.cpp
           zip -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-x64.zip ./build/bin/*
 
       - 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-ubuntu-x64.zip
@@ -784,12 +790,14 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           Copy-Item LICENSE .\build\bin\Release\llama.cpp.txt
           Copy-Item .\examples\run\linenoise.cpp\LICENSE .\build\bin\Release\linenoise.cpp.txt
           7z a llama-${{ 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 }}.zip
@@ -932,16 +940,19 @@ jobs:
 
       - name: Pack artifacts
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           7z a llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}-cu${{ matrix.cuda }}-x64.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-${{ matrix.build }}-cu${{ matrix.cuda }}-x64.zip
           name: llama-bin-win-cu${{ matrix.cuda }}-x64.zip
 
       - name: Copy and pack Cuda runtime
+        if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
         run: |
           echo "Cuda install location: ${{ env.CUDA_PATH }}"
           $dst='.\build\bin\cudart\'
@@ -950,6 +961,7 @@ jobs:
           7z a cudart-llama-bin-win-cu${{ matrix.cuda }}-x64.zip $dst\*
 
       - name: Upload Cuda runtime
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         uses: actions/upload-artifact@v4
         with:
           path: cudart-llama-bin-win-cu${{ matrix.cuda }}-x64.zip
@@ -996,6 +1008,7 @@ jobs:
 
       - name: Build the release package
         id: pack_artifacts
+        if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }}
         run: |
           echo "cp oneAPI running time dll files in ${{ env.ONEAPI_ROOT }} to ./build/bin"
 
@@ -1020,6 +1033,7 @@ jobs:
           7z a llama-${{ steps.tag.outputs.name }}-bin-win-sycl-x64.zip ./build/bin/*
 
       - name: Upload the release package
+        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-sycl-x64.zip