Merge pull request #385 from leofang/component_11 #135
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: "CI: Build and test" | |
| concurrency: | |
| group: ${{ github.workflow }}-${{ | |
| github.ref_name == '11.8.x' && format('ci-main-build-test-{0}', github.run_id) || | |
| format('ci-pr-build-test-on-{0}-against-branch-{1}', github.event_name, github.ref_name) | |
| }} | |
| cancel-in-progress: true | |
| on: | |
| push: | |
| branches: | |
| - "pull-request/[0-9]+" | |
| - "11.8.x" | |
| jobs: | |
| build: | |
| strategy: | |
| fail-fast: false | |
| matrix: | |
| host-platform: | |
| - linux-64 | |
| - linux-aarch64 | |
| - win-64 | |
| python-version: | |
| - "3.13" | |
| - "3.12" | |
| - "3.11" | |
| - "3.10" | |
| - "3.9" | |
| cuda-version: | |
| # Note: this is for build-time only. | |
| - "11.8.0" | |
| name: Build (${{ matrix.host-platform }}, Python ${{ matrix.python-version }}, CUDA ${{ matrix.cuda-version }}) | |
| if: ${{ github.repository_owner == 'nvidia' }} | |
| permissions: | |
| id-token: write # This is required for configure-aws-credentials | |
| contents: read # This is required for actions/checkout | |
| runs-on: ${{ (matrix.host-platform == 'linux-64' && 'linux-amd64-cpu8') || | |
| (matrix.host-platform == 'linux-aarch64' && 'linux-arm64-cpu8') || | |
| (matrix.host-platform == 'win-64' && 'windows-2019') }} | |
| # (matrix.host-platform == 'win-64' && 'windows-amd64-cpu8') }} | |
| outputs: | |
| BUILD_CTK_VER: ${{ steps.pass_env.outputs.CUDA_VERSION }} | |
| defaults: | |
| run: | |
| shell: bash --noprofile --norc -xeuo pipefail {0} | |
| steps: | |
| - name: Checkout ${{ github.event.repository.name }} | |
| uses: actions/checkout@v4 | |
| with: | |
| fetch-depth: 0 | |
| ref: ${{ github.head_ref || github.ref_name }} | |
| # WAR: setup-python is not relocatable... | |
| # see https://github.com/actions/setup-python/issues/871 | |
| - name: Set up Python ${{ matrix.python-version }} | |
| if: ${{ startsWith(matrix.host-platform, 'linux') }} | |
| id: setup-python | |
| uses: actions/setup-python@v5 | |
| with: | |
| python-version: "3.12" | |
| - name: Set up MSVC | |
| if: ${{ startsWith(matrix.host-platform, 'win') }} | |
| uses: ilammy/msvc-dev-cmd@v1 | |
| - name: Set environment variables | |
| run: | | |
| PYTHON_VERSION_FORMATTED=$(echo '${{ matrix.python-version }}' | tr -d '.') | |
| if [[ "${{ matrix.host-platform }}" == linux* ]]; then | |
| CIBW_BUILD="cp${PYTHON_VERSION_FORMATTED}-manylinux*" | |
| REPO_DIR=$(pwd) | |
| elif [[ "${{ matrix.host-platform }}" == win* ]]; then | |
| CIBW_BUILD="cp${PYTHON_VERSION_FORMATTED}-win_amd64" | |
| PWD=$(pwd) | |
| REPO_DIR=$(cygpath -w $PWD) | |
| fi | |
| echo "PARALLEL_LEVEL=$(nproc)" >> $GITHUB_ENV | |
| echo "CUDA_BINDINGS_ARTIFACT_NAME=cuda-bindings-python${PYTHON_VERSION_FORMATTED}-cuda${{ matrix.cuda-version }}-${{ matrix.host-platform }}-${{ github.sha }}" >> $GITHUB_ENV | |
| echo "CUDA_BINDINGS_ARTIFACTS_DIR=$(realpath "$REPO_DIR/cuda_bindings/dist")" >> $GITHUB_ENV | |
| echo "CIBW_BUILD=${CIBW_BUILD}" >> $GITHUB_ENV | |
| - name: Dump environment | |
| run: | | |
| env | |
| - name: Set up mini CTK | |
| uses: ./.github/actions/fetch_ctk | |
| continue-on-error: false | |
| with: | |
| host-platform: ${{ matrix.host-platform }} | |
| cuda-version: ${{ matrix.cuda-version }} | |
| - name: Build cuda.bindings wheel | |
| uses: pypa/[email protected] | |
| env: | |
| CIBW_BUILD: ${{ env.CIBW_BUILD }} | |
| CIBW_ARCHS_LINUX: "native" | |
| CIBW_BUILD_VERBOSITY: 1 | |
| # CIBW mounts the host filesystem under /host | |
| CIBW_ENVIRONMENT_LINUX: > | |
| CUDA_PATH=/host/${{ env.CUDA_PATH }} | |
| PARALLEL_LEVEL=${{ env.PARALLEL_LEVEL }} | |
| CIBW_ENVIRONMENT_WINDOWS: > | |
| CUDA_HOME="$(cygpath -w ${{ env.CUDA_PATH }})" | |
| # PARALLEL_LEVEL=${{ env.PARALLEL_LEVEL }} | |
| with: | |
| package-dir: ./cuda_bindings/ | |
| output-dir: ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }} | |
| - name: List the cuda.bindings artifacts directory | |
| run: | | |
| if [[ "${{ matrix.host-platform }}" == win* ]]; then | |
| export CHOWN=chown | |
| else | |
| export CHOWN="sudo chown" | |
| fi | |
| $CHOWN -R $(whoami) ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }} | |
| ls -lahR ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }} | |
| # TODO: enable this after NVIDIA/cuda-python#297 is resolved | |
| # - name: Check cuda.bindings wheel | |
| # run: | | |
| # pip install twine | |
| # twine check ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }}/*.whl | |
| - name: Upload cuda.bindings build artifacts | |
| uses: actions/upload-artifact@v4 | |
| with: | |
| name: ${{ env.CUDA_BINDINGS_ARTIFACT_NAME }} | |
| path: ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }}/*.whl | |
| if-no-files-found: error | |
| # upload-artifact's "overwrite: true" option has a race condition among parallel | |
| # jobs, so we let job 0 do the work | |
| - name: Build and check cuda-python wheel | |
| if: ${{ strategy.job-index == 0 }} | |
| run: | | |
| pushd cuda_python | |
| pip wheel -v --no-deps . | |
| pip install twine # TODO: remove me (see above) | |
| twine check *.whl | |
| popd | |
| - name: List the cuda-python artifacts directory | |
| if: ${{ strategy.job-index == 0 }} | |
| run: | | |
| if [[ "${{ matrix.host-platform }}" == win* ]]; then | |
| export CHOWN=chown | |
| else | |
| export CHOWN="sudo chown" | |
| fi | |
| $CHOWN -R $(whoami) cuda_python/*.whl | |
| ls -lahR cuda_python | |
| - name: Upload cuda-python build artifacts | |
| if: ${{ strategy.job-index == 0 }} | |
| uses: actions/upload-artifact@v4 | |
| with: | |
| name: cuda-python-wheel | |
| path: cuda_python/*.whl | |
| if-no-files-found: error | |
| - name: Pass environment variables to the next runner | |
| id: pass_env | |
| run: | | |
| echo "CUDA_VERSION=${{ matrix.cuda-version }}" >> $GITHUB_OUTPUT | |
| test: | |
| strategy: | |
| fail-fast: false | |
| # TODO: add driver version here | |
| matrix: | |
| host-platform: | |
| - linux-64 | |
| - linux-aarch64 | |
| # TODO: enable testing once win-64 GPU runners are up | |
| # - win-64 | |
| python-version: | |
| - "3.13" | |
| - "3.12" | |
| - "3.11" | |
| - "3.10" | |
| - "3.9" | |
| cuda-version: | |
| # Note: this is for test-time only. | |
| - "11.8.0" | |
| local-ctk: | |
| - 1 # use mini CTK | |
| - 0 # use CTK wheels | |
| runner: | |
| - default | |
| include: | |
| - host-platform: linux-64 | |
| python-version: "3.12" | |
| cuda-version: "11.8.0" | |
| runner: H100 | |
| name: Test (${{ matrix.host-platform }}, Python ${{ matrix.python-version }}, CUDA ${{ matrix.cuda-version }}, Runner ${{ matrix.runner }}, ${{ (matrix.local-ctk == '1' && 'local CTK') || 'CTK wheels' }}) | |
| # The build stage could fail but we want the CI to keep moving. | |
| if: ${{ github.repository_owner == 'nvidia' && always() }} | |
| permissions: | |
| id-token: write # This is required for configure-aws-credentials | |
| contents: read # This is required for actions/checkout | |
| runs-on: ${{ (matrix.runner == 'default' && matrix.host-platform == 'linux-64' && 'linux-amd64-gpu-v100-latest-1') || | |
| (matrix.runner == 'default' && matrix.host-platform == 'linux-aarch64' && 'linux-arm64-gpu-a100-latest-1') || | |
| (matrix.runner == 'H100' && 'linux-amd64-gpu-h100-latest-1-testing') }} | |
| # Our self-hosted runners require a container | |
| # TODO: use a different (nvidia?) container | |
| container: | |
| options: -u root --security-opt seccomp=unconfined --shm-size 16g | |
| image: ubuntu:22.04 | |
| env: | |
| NVIDIA_VISIBLE_DEVICES: ${{ env.NVIDIA_VISIBLE_DEVICES }} | |
| needs: | |
| - build | |
| defaults: | |
| run: | |
| shell: bash --noprofile --norc -xeuo pipefail {0} | |
| steps: | |
| - name: Ensure GPU is working | |
| run: nvidia-smi | |
| - name: Checkout ${{ github.event.repository.name }} | |
| uses: actions/checkout@v4 | |
| with: | |
| fetch-depth: 0 | |
| ref: ${{ github.head_ref || github.ref_name }} | |
| - name: Set environment variables | |
| run: | | |
| PYTHON_VERSION_FORMATTED=$(echo '${{ matrix.python-version }}' | tr -d '.') | |
| if [[ "${{ matrix.host-platform }}" == linux* ]]; then | |
| REPO_DIR=$(pwd) | |
| elif [[ "${{ matrix.host-platform }}" == win* ]]; then | |
| PWD=$(pwd) | |
| REPO_DIR=$(cygpath -w $PWD) | |
| fi | |
| BUILD_CUDA_MAJOR="$(cut -d '.' -f 1 <<< ${{ needs.build.outputs.BUILD_CTK_VER }})" | |
| TEST_CUDA_MAJOR="$(cut -d '.' -f 1 <<< ${{ matrix.cuda-version }})" | |
| if [[ $BUILD_CUDA_MAJOR != $TEST_CUDA_MAJOR ]]; then | |
| SKIP_CUDA_BINDINGS_TEST=1 | |
| else | |
| SKIP_CUDA_BINDINGS_TEST=0 | |
| fi | |
| # make outputs from the previous job as env vars | |
| echo "CUDA_BINDINGS_ARTIFACT_NAME=cuda-bindings-python${PYTHON_VERSION_FORMATTED}-cuda${{ needs.build.outputs.BUILD_CTK_VER }}-${{ matrix.host-platform }}-${{ github.sha }}" >> $GITHUB_ENV | |
| echo "CUDA_BINDINGS_ARTIFACTS_DIR=$(realpath "$REPO_DIR/cuda_bindings/dist")" >> $GITHUB_ENV | |
| echo "SKIP_CUDA_BINDINGS_TEST=${SKIP_CUDA_BINDINGS_TEST}" >> $GITHUB_ENV | |
| - name: Download cuda-python build artifacts | |
| uses: actions/download-artifact@v4 | |
| with: | |
| name: cuda-python-wheel | |
| path: . | |
| - name: Display structure of downloaded cuda-python artifacts | |
| run: | | |
| pwd | |
| ls -lahR . | |
| - name: Download cuda.bindings build artifacts | |
| uses: actions/download-artifact@v4 | |
| with: | |
| name: ${{ env.CUDA_BINDINGS_ARTIFACT_NAME }} | |
| path: ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }} | |
| - name: Display structure of downloaded cuda.bindings artifacts | |
| run: | | |
| pwd | |
| ls -lahR $CUDA_BINDINGS_ARTIFACTS_DIR | |
| - name: Set up Python ${{ matrix.python-version }} | |
| uses: actions/setup-python@v5 | |
| with: | |
| python-version: ${{ matrix.python-version }} | |
| env: | |
| # we use self-hosted runners on which setup-python behaves weirdly... | |
| AGENT_TOOLSDIRECTORY: "/opt/hostedtoolcache" | |
| - name: Set up mini CTK | |
| if: ${{ matrix.local-ctk == '1' }} | |
| uses: ./.github/actions/fetch_ctk | |
| continue-on-error: false | |
| with: | |
| host-platform: ${{ matrix.host-platform }} | |
| cuda-version: ${{ matrix.cuda-version }} | |
| - name: Run cuda.bindings tests | |
| if: ${{ env.SKIP_CUDA_BINDINGS_TEST == '0' }} | |
| run: | | |
| pushd "${CUDA_BINDINGS_ARTIFACTS_DIR}" | |
| if [[ "${{ matrix.local-ctk }}" == 1 ]]; then | |
| ls $CUDA_PATH | |
| pip install *.whl | |
| else | |
| pip install $(ls *.whl)[all] | |
| fi | |
| popd | |
| pushd ./cuda_bindings | |
| pip install -r requirements.txt | |
| pytest -rxXs tests/ | |
| # It is a bit convoluted to run the Cython tests against CTK wheels, | |
| # so let's just skip them. | |
| if [[ "${{ matrix.local-ctk }}" == 1 ]]; then | |
| if [[ "${{ matrix.host-platform }}" == linux* ]]; then | |
| # cython tests require gcc | |
| apt install -y build-essential | |
| bash tests_cython/build_tests.sh | |
| elif [[ "${{ matrix.host-platform }}" == win* ]]; then | |
| # TODO: enable this once win-64 runners are up | |
| exit 1 | |
| fi | |
| pytest -rxXs tests_cython | |
| fi | |
| popd | |
| - name: Ensure cuda-python installable | |
| run: | | |
| pip install cuda_python*.whl | |
| checks: | |
| name: Check job status | |
| permissions: | |
| checks: read | |
| needs: | |
| - build | |
| - test | |
| secrets: inherit | |
| uses: | |
| ./.github/workflows/status-check.yml |