-
Notifications
You must be signed in to change notification settings - Fork 43
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Add GHA build-hpc running on ECMWF GPU partition
- Loading branch information
1 parent
af64837
commit 04a0a25
Showing
1 changed file
with
176 additions
and
0 deletions.
There are no files selected for viewing
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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,176 @@ | ||
| name: build-hpc | ||
|
|
||
| # Controls when the action will run | ||
| on: | ||
|
|
||
| # Trigger the workflow on all pushes to main and develop, except on tag creation | ||
| push: | ||
| branches: | ||
| - main | ||
| - develop | ||
| - ci | ||
| tags-ignore: | ||
| - '**' | ||
|
|
||
| # Trigger the workflow on all pull requests | ||
| pull_request: ~ | ||
|
|
||
| # Allow workflow to be dispatched on demand | ||
| workflow_dispatch: ~ | ||
|
|
||
| # Trigger after public PR approved for CI | ||
| pull_request_target: | ||
| types: [labeled] | ||
|
|
||
| env: | ||
| ATLAS_TOOLS: ${{ github.workspace }}/tools | ||
| CTEST_PARALLEL_LEVEL: 1 | ||
| CACHE_SUFFIX: v1 # Increase to force new cache to be created | ||
|
|
||
| jobs: | ||
| ci-hpc: | ||
| name: ci-hpc | ||
| if: ${{ !github.event.pull_request.head.repo.fork && github.event.action != 'labeled' || github.event.label.name == 'approved-for-ci' }} | ||
|
|
||
| strategy: | ||
| fail-fast: false # false: try to complete all jobs | ||
|
|
||
| matrix: | ||
| name: | ||
| - ac-gpu nvhpc | ||
|
|
||
| include: | ||
| - name: ac-gpu nvhpc | ||
| site: ac-batch | ||
| troika_user_secret: HPC_CI_GPU_SSH_USER | ||
| sbatch_options: | | ||
| #SBATCH --time=00:20:00 | ||
| #SBATCH --nodes=1 | ||
| #SBATCH --ntasks=4 | ||
| #SBATCH --cpus-per-task=32 | ||
| #SBATCH --gpus-per-task=1 | ||
| #SBATCH --mem=200G | ||
| #SBATCH --qos=dg | ||
| modules: | ||
| - cmake | ||
| - ninja | ||
| - prgenv/nvidia | ||
| - hpcx-openmpi/2.14.0-cuda | ||
| - fftw | ||
| - qhull | ||
| - eigen | ||
| cmake_options: -DMPI_SLOTS=4 | ||
|
|
||
| runs-on: [self-hosted, linux, hpc] | ||
| env: | ||
| GH_TOKEN: ${{ github.token }} | ||
| steps: | ||
| - uses: ecmwf-actions/reusable-workflows/ci-hpc-generic@v2 | ||
| with: | ||
| site: ${{ matrix.site }} | ||
| troika_user: ${{ secrets[matrix.troika_user_secret] }} | ||
| sbatch_options: ${{ matrix.sbatch_options }} | ||
| output_dir: ${{ matrix.output_dir || '' }} | ||
| workdir: ${{ matrix.workdir || '' }} | ||
| template_data: | | ||
| cmake_options: | ||
| - -DENABLE_MPI=ON | ||
| - -DENABLE_ACC=ON | ||
| - -DENABLE_CUDA=ON | ||
| - -DSKIP_TEST_atlas_test_field_foreach=TRUE | ||
| - ${{ matrix.cmake_options || '' }} | ||
| ctest_options: ${{ matrix.ctest_options || '' }} | ||
| dependencies: | ||
| ecmwf/ecbuild: | ||
| version: develop | ||
| ecmwf/eckit: | ||
| version: develop | ||
| cmake_options: | ||
| - -DENABLE_MPI=ON | ||
| - -DENABLE_CUDA=OFF | ||
| - -DENABLE_TESTS=OFF | ||
| - -DENABLE_ECKIT_SQL=OFF | ||
| - -DENABLE_ECKIT_CMD=OFF | ||
| - -DENABLE_AIO=OFF | ||
| - -DENABLE_WARNINGS=OFF | ||
| - ${{ matrix.cmake_options || '' }} | ||
| ecmwf/fckit: | ||
| version: develop | ||
| cmake_options: | ||
| - -DENABLE_TESTS=OFF | ||
| - ${{ matrix.cmake_options || '' }} | ||
| ecmwf-ifs/fiat: | ||
| version: develop | ||
| cmake_options: | ||
| - -DENABLE_MPI=ON | ||
| - -DENABLE_TESTS=OFF | ||
| - ${{ matrix.cmake_options || '' }} | ||
| ecmwf-ifs/ectrans: | ||
| version: develop | ||
| cmake_options: | ||
| - -DENABLE_MPI=ON | ||
| - -DENABLE_ACC=ON | ||
| - -DENABLE_GPU=ON | ||
| - -DENABLE_TESTS=OFF | ||
| - ${{ matrix.cmake_options || '' }} | ||
| template: | | ||
| set +x | ||
| {% for module in "${{ join(matrix.modules, ',') }}".split(',') %} | ||
| module load {{module}} | ||
| {% endfor %} | ||
| echo "+ module list" | ||
| module list | ||
| set -x | ||
| BASEDIR=$PWD | ||
| export CMAKE_TEST_LAUNCHER="srun;-n;1" | ||
| export CMAKE_PREFIX_PATH=$BASEDIR/install:$CMAKE_PREFIX_PATH | ||
| {% for repo_name, options in dependencies.items() %} | ||
| name=$(basename {{repo_name}}) | ||
| mkdir -p $name | ||
| pushd $name | ||
| git init | ||
| git remote add origin ${{ github.server_url }}/{{repo_name}} | ||
| git fetch origin {{options['version']}} | ||
| git reset --hard FETCH_HEAD | ||
| cmake -G Ninja -S . -B build \ | ||
| {{ options['cmake_options']|join(' ') }} | ||
| start=`date +%s` | ||
| cmake --build build | ||
| end=`date +%s` | ||
| runtime=$((end-start)) | ||
| echo "Build $name took $runtime seconds" | ||
| cmake --install build --prefix $BASEDIR/install/$name | ||
| export PATH=$BASEDIR/install/$name/bin:$PATH | ||
| popd | ||
| {% endfor %} | ||
| REPO=${{ github.event.pull_request.head.repo.full_name || github.repository }} | ||
| SHA=${{ github.event.pull_request.head.sha || github.sha }} | ||
| mkdir -p $REPO | ||
| pushd $REPO | ||
| git init | ||
| git remote add origin ${{ github.server_url }}/$REPO | ||
| git fetch origin $SHA | ||
| git reset --hard FETCH_HEAD | ||
| popd | ||
| cmake -G Ninja -S $REPO -B build \ | ||
| {{ cmake_options|join(' ') }} | ||
| start=`date +%s` | ||
| cmake --build build | ||
| end=`date +%s` | ||
| runtime=$((end-start)) | ||
| echo "Build $name took $runtime seconds" | ||
| export ATLAS_FINALISES_MPI=1 | ||
| ctest --test-dir build --output-on-failure {{ ctest_options }} | ||
| cmake --install build --prefix $BASEDIR/install/$REPO | ||
| export PATH=$BASEDIR/install/$REPO/bin:$PATH | ||
| atlas --info | ||
| {% for repo_name in dependencies.keys() %} | ||
| name=$(basename {{repo_name}}) | ||
| rm -r $name | ||
| {% endfor %} | ||
| rm -r $REPO | ||
| rm -r build | ||
| rm -r $BASEDIR/install |