Add a multi-node CI job for MPI Shared Memory

Author: hmenke

.github/workflows/build_multi_node.yml

@@ -0,0 +1,88 @@
+name: build
+
+on:
+  push:
+    branches: [ unstable ]
+  pull_request:
+    branches: [ unstable ]
+  workflow_call:
+  workflow_dispatch:
+
+env:
+  CMAKE_C_COMPILER_LAUNCHER: ccache
+  CMAKE_CXX_COMPILER_LAUNCHER: ccache
+  CCACHE_COMPILERCHECK: content
+  CCACHE_BASEDIR: ${{ github.workspace }}
+  CCACHE_DIR: ${{ github.workspace }}/.ccache
+  CCACHE_MAXSIZE: 500M
+  CCACHE_SLOPPINESS: pch_defines,time_macros,include_file_mtime,include_file_ctime
+  CCACHE_COMPRESS: "1"
+  CCACHE_COMPRESSLEVEL: "1"
+
+jobs:
+  build_multi_node:
+
+    strategy:
+      fail-fast: false
+
+    runs-on: ubuntu-24.04
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - uses: actions/cache/restore@v4
+      with:
+        path: ${{ env.CCACHE_DIR }}
+        key: ccache-${{ matrix.os }}-${{ matrix.cc }}-${{ github.run_id }}
+        restore-keys:
+          ccache-${{ matrix.os }}-${{ matrix.cc }}-
+
+    - name: Install ubuntu dependencies
+      run: >
+        sudo apt-get update && sudo apt-get install ccache
+
+    - name: Build and start Docker Compose
+      run: |
+        docker compose build
+        docker compose up -d
+      working-directory: .github/workflows/docker
+
+    - name: Compile MPI inside the container
+      run: |
+        docker exec -t -w ${{ github.workspace }} docker-vm-1 /bin/bash -euxc '
+        cmake -S . -B build -DCMAKE_INSTALL_PREFIX=$HOME/install -DBuild_Documentation=Off
+        cmake --build build/ -j2
+        '
+
+    - name: Run tests inside the container
+      run: |
+        docker exec -t -w ${{ github.workspace }} docker-vm-1 /bin/bash -euxc '
+        export CTEST_OUTPUT_ON_FAILURE=1
+        cmake --build build/ --target test
+        '
+
+    - name: Run tests inside the container
+      run: |
+        docker exec -t -w ${{ github.workspace }} docker-vm-1 /bin/bash -euxc '
+        cat <<EOF | tee hostfile
+        docker-vm-1 slots=3
+        docker-vm-2 slots=1
+        docker-vm-3 slots=2
+        EOF
+
+        # Test the communication
+        mpirun -hostfile ./hostfile /bin/bash -c "env | grep \"^OMPI_COMM_.*_RANK\""
+
+        # Run the actual test
+        mpirun -hostfile ./hostfile build/test/c++/mpi_window
+        '
+
+    - name: ccache statistics
+      if: always()
+      run: ccache -sv
+
+    - uses: actions/cache/save@v4
+      if: always()
+      with:
+        path: ${{ env.CCACHE_DIR }}
+        key: ccache-${{ matrix.os }}-${{ matrix.cc }}-${{ github.run_id }}

.github/workflows/docker/Dockerfile

@@ -0,0 +1,27 @@
+FROM ubuntu:24.04
+ENV DEBIAN_FRONTEND=noninteractive
+RUN apt-get update \
+ && apt-get install --no-install-recommends -y \
+    ca-certificates \
+    build-essential \
+    ccache \
+    cmake \
+    g++ \
+    git \
+    libopenmpi-dev \
+    openmpi-bin \
+    openssh-server \
+ && rm -rf /var/cache/apt /var/lib/apt/lists
+RUN passwd -d root \
+ && mkdir -pm0755 /run/sshd \
+ && echo "PermitRootLogin yes" >> /etc/ssh/sshd_config.d/99-insecure.conf \
+ && echo "PasswordAuthentication yes" >> /etc/ssh/sshd_config.d/99-insecure.conf \
+ && echo "PermitEmptyPasswords yes" >> /etc/ssh/sshd_config.d/99-insecure.conf \
+ && echo "StrictHostKeyChecking no" >> /etc/ssh/ssh_config.d/99-insecure.conf \
+ && echo "LogLevel ERROR" >> /etc/ssh/ssh_config.d/99-insecure.conf
+ENV OMPI_ALLOW_RUN_AS_ROOT=1
+ENV OMPI_ALLOW_RUN_AS_ROOT_CONFIRM=1
+ENV OMPI_MCA_btl_vader_single_copy_mechanism=none
+ENV OMPI_MCA_osc=sm,pt2pt
+ENV OMPI_MCA_rmaps_base_oversubscribe=yes
+CMD ["/usr/sbin/sshd", "-D"]

.github/workflows/docker/docker-compose.yml

@@ -0,0 +1,20 @@
+services:
+  vm:
+    build: .
+    deploy:
+      mode: replicated
+      replicas: 3
+    environment:
+      CMAKE_C_COMPILER_LAUNCHER: ${CMAKE_C_COMPILER_LAUNCHER}
+      CMAKE_CXX_COMPILER_LAUNCHER: ${CMAKE_CXX_COMPILER_LAUNCHER}
+      CCACHE_COMPILERCHECK: ${CCACHE_COMPILERCHECK}
+      CCACHE_BASEDIR: ${CCACHE_BASEDIR}
+      CCACHE_DIR: ${CCACHE_DIR}
+      CCACHE_MAXSIZE: ${CCACHE_MAXSIZE}
+      CCACHE_SLOPPINESS: ${CCACHE_SLOPPINESS}
+      CCACHE_COMPRESS: ${CCACHE_COMPRESS}
+      CCACHE_COMPRESSLEVEL: ${CCACHE_COMPRESSLEVEL}
+    stdin_open: true
+    tty: true
+    volumes:
+      - ${CCACHE_BASEDIR}:${CCACHE_BASEDIR}:z