gpu_nvhpc.yml

name: Linux NVHPC GPU
# triggered events (push, pull_request) for the master branch
on:
  pull_request:
    branches: [ master ]
    types: [ labeled, opened, synchronize, reopened ]
  workflow_dispatch:

#defaults:
#  run:
#    shell: bash -leo pipefail {0}

# A workflow run is made up of one or more jobs that can run sequentially or in parallel
jobs:

  ubuntu_build:
    if: contains(github.event.pull_request.labels.*.name, 'GPU-CI')

    name: Ubuntu NVHPC GPU Build
    # Run on self-hosted
    runs-on: self-hosted

    steps:

      ## Install Lmod
      #- name: Install Lmod
      #  run: |
      #    sudo apt-get update -y
      #    sudo apt-get install -y lmod
      #    echo "source /usr/share/lmod/lmod/init/bash" >> ~/.bash_profile
      #    source /usr/share/lmod/lmod/init/bash
      #    module list
      #
      ## Install NVIDIA HPC SDK
      #- name: Install NVIDIA HPC SDK
      #  run: |
      #    curl https://developer.download.nvidia.com/hpc-sdk/ubuntu/DEB-GPG-KEY-NVIDIA-HPC-SDK | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-hpcsdk-archive-keyring.gpg
      #    echo 'deb [signed-by=/usr/share/keyrings/nvidia-hpcsdk-archive-keyring.gpg] https://developer.download.nvidia.com/hpc-sdk/ubuntu/amd64 /' | sudo tee /etc/apt/sources.list.d/nvhpc.list
      #    sudo apt-get update -y
      #    sudo apt-get install -y nvhpc-24-7

      # Remove label
      #- name: Remove GPU-CI label
      #  - uses: actions-ecosystem/action-remove-labels@v1
      #  with:
      #    labels: GPU-CI

      # Check location of installed NVHPC compilers
      - name: Check compiler install
        run: |
          source /usr/share/lmod/lmod/init/bash
          module use /opt/nvidia/hpc_sdk/modulefiles
          module load nvhpc
          which nvc
          which nvfortran

      # Test debug mode
      # Turn this off because the compiler hangs while building in debug mode
      #- name: Build gf debug
      #  run: |
      #    source /usr/share/lmod/lmod/init/bash
      #    module use /opt/nvidia/hpc_sdk/modulefiles
      #    module load nvhpc
      #    cd ref
      #    rm -rf build
      #    mkdir build
      #    cd build
      #    #export OMP_NUM_THREADS=4
      #    cmake -DCMAKE_BUILD_TYPE=debug -DENABLE_GPU=on ..
      #    make VERBOSE=1
      #    ctest --output-on-failure

      # Test release mode
      - name: Build gf release
        run: |
          source /usr/share/lmod/lmod/init/bash
          module use /opt/nvidia/hpc_sdk/modulefiles
          module load nvhpc
          cd ref
          rm -rf build
          mkdir build
          cd build
          #export OMP_NUM_THREADS=4
          cmake -DCMAKE_BUILD_TYPE=release -DENABLE_GPU=on ..
          make VERBOSE=1
          ulimit -s hard
          ctest --output-on-failure

      # Debug session for failures
      -
        name: Debug session
        if: ${{ failure() }}
        uses: mxschmitt/action-tmate@v3
        timeout-minutes: 60
        with:
          limit-access-to-actor: true