[feat] implement async llm engine for python wrapper (#172) #2
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: build and test | |
| on: | |
| push: | |
| paths-ignore: | |
| - ".github/docker.yml" | |
| - "docs/**" | |
| - "tools/**" | |
| - "gateway/**" | |
| - "scripts/**" | |
| - "**/*.md" | |
| - "**/*.txt" | |
| - "**/*.sh" | |
| - "Dockerfile" | |
| - "Dockerfile.*" | |
| branches: | |
| - main | |
| pull_request: | |
| paths-ignore: | |
| - ".github/docker.yml" | |
| - "docs/**" | |
| - "tools/**" | |
| - "gateway/**" | |
| - "scripts/**" | |
| - "**/*.md" | |
| - "**/*.txt" | |
| - "**/*.sh" | |
| - "Dockerfile" | |
| - "Dockerfile.*" | |
| branches: | |
| - main | |
| jobs: | |
| linux-gcc: | |
| strategy: | |
| matrix: | |
| os: [ubuntu-22.04] | |
| # Customize the CMake build type here (Release, Debug, RelWithDebInfo, etc.) | |
| build_type: [Release] | |
| runs-on: [self-hosted, linux, x64, build] | |
| env: | |
| BUILD_TYPE: ${{ matrix.build_type }} | |
| # Tells vcpkg where binary packages are stored. | |
| VCPKG_DEFAULT_BINARY_CACHE: ${{ github.workspace }}/../../_vcpkg/bincache | |
| # Tells ccache where to store its cache. | |
| CCACHE_DIR: ${{ github.workspace }}/../../_ccache | |
| # Tells where to store the dependencies. | |
| DEPENDENCES_ROOT: ${{ github.workspace }}/../../_deps | |
| steps: | |
| - name: Install toolkits | |
| run: | | |
| sudo apt-get install -y build-essential ninja-build bison gcc-12 g++-12 libunwind-dev python3-dev libboost-all-dev ccache | |
| sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-12 10 --slave /usr/bin/g++ g++ /usr/bin/g++-12 | |
| - name: Show gcc version | |
| run: gcc --version | |
| - name: Checkout | |
| uses: actions/checkout@v3 | |
| with: | |
| submodules: recursive | |
| - name: Zero out ccache statistics | |
| run: ccache -z | |
| # Restore vcpkg bincache from the newest or latest cache available. | |
| # TODO: use a distribute storage for bincache by following the instruction here: | |
| # https://learn.microsoft.com/en-us/vcpkg/users/binarycaching#aws | |
| # - name: Restore vcpkg and its artifacts. | |
| # uses: actions/cache@v3 | |
| # with: | |
| # path: ${{ env.VCPKG_DEFAULT_BINARY_CACHE }} | |
| # key: | | |
| # scalellm-${{ matrix.os }}-gcc-x64-linux-${{env.BUILD_TYPE}}-${{ hashFiles( 'vcpkg.json' ) }} | |
| # restore-keys: | | |
| # scalellm-${{ matrix.os }}-gcc-x64-linux-${{env.BUILD_TYPE}} | |
| # - name: Restore ccache and its artifacts. | |
| # uses: actions/cache@v3 | |
| # with: | |
| # path: ${{ env.CCACHE_DIR }} | |
| # key: | | |
| # scalellm-${{ matrix.os }}-gcc-x64-linux-ccache-${{env.BUILD_TYPE}}-${{ github.sha }} | |
| # restore-keys: | | |
| # scalellm-${{ matrix.os }}-gcc-x64-linux-ccache-${{env.BUILD_TYPE}} | |
| - name: CMake Build in the 'build' subdirectory | |
| run: | | |
| cmake -G Ninja -S . -B build -DCMAKE_CUDA_ARCHITECTURES=80 -DTORCH_CUDA_ARCH_LIST=8.0 -DUSE_CCACHE=ON | |
| cmake --build build --config ${{env.BUILD_TYPE}} --target all | |
| # - name: benchmark | |
| # if: ${{ env.BUILD_TYPE == 'Release' }} | |
| # run: ./build/src/benchmark/micro_benchmark | |
| - name: Run unit tests | |
| working-directory: ${{github.workspace}}/build | |
| # Execute tests defined by the CMake configuration. | |
| run: ctest --output-on-failure -C ${{env.BUILD_TYPE}} | |
| - name: Show ccache statistics | |
| run: ccache -s |