Intel oneAPI GitHub Actions with MKL and MPI (C, C++, Fortran) and CMake

github-actions-oneapi.yml

jobs:

  linux-intel-oneapi:
    runs-on: ubuntu-latest
    env:
      CC: icx
      CXX: icpx
      FC: ifort

    steps:
    - name: Intel Apt repository
      timeout-minutes: 1
      run: |
        wget https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS-2023.PUB
        sudo apt-key add GPG-PUB-KEY-INTEL-SW-PRODUCTS-2023.PUB
        rm GPG-PUB-KEY-INTEL-SW-PRODUCTS-2023.PUB
        echo "deb https://apt.repos.intel.com/oneapi all main" | sudo tee /etc/apt/sources.list.d/oneAPI.list
        sudo apt-get update
        
    - name: Install Intel oneAPI
      timeout-minutes: 5
      run: sudo apt-get install intel-oneapi-compiler-fortran intel-oneapi-compiler-dpcpp-cpp-and-cpp-classic intel-oneapi-mpi intel-oneapi-mpi-devel intel-oneapi-mkl ninja-build

    - name: checkout project code
      uses: actions/checkout@v2

    - name: Setup Intel oneAPI environment
      run: |
        source /opt/intel/oneapi/setvars.sh
        printenv >> $GITHUB_ENV
        
    - name: CMake Configure
      run: cmake -B build

    - name: CMake build
      run: cmake --build build

    - name: CMake test
      run: ctest --test-dir build

Set env: CC: etc workflow-wide so that ExternalProjects get the Intel compiler too

Can you update this with caching the installation step?

https://github.com/scivision/fortran-filesystem/blob/main/.github/workflows/oneapi-linux.yml
This uses auxiliary scripts and cache still takes a couple minutes to setup. I didn't switch to cached in other projects as a result of not much time savings and additional setup hassle

I was using the exact same Install Intel oneAPI step (without the ninja_build) and the apt installation takes almost 3 and half minutes, whereas the cache setup took 50 sec.