llama.cpp/.github/workflows/ci-self-hosted-vulkan.yml

name: CI (self-hosted Vulkan backend)

on:
  workflow_dispatch: # allows manual triggering
  push:
    branches:
      - master
    paths: [
      '.github/workflows/ci-self-hosted-vulkan.yml',
      'ci/run.sh',
      '**/CMakeLists.txt',
      '**/.cmake',
      '**/*.h',
      '**/*.hpp',
      '**/*.c',
      '**/*.cpp',
      '**/*.comp',
      '**/*.glsl'
    ]

  pull_request:
    types: [opened, synchronize, reopened]
    paths: [
      '.github/workflows/ci-self-hosted-vulkan.yml',
      'ci/run.sh',
      '**/CMakeLists.txt',
      '**/.cmake',
      'ggml/src/*',
      'ggml/src/ggml-cpu/**',
      'ggml/src/ggml-vulkan/**'
    ]

concurrency:
  group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
  cancel-in-progress: true

env:
  # note: this is dud token to avoid rate limiting (https://github.com/ggml-org/llama.cpp/pull/25706#issuecomment-4979941302)
  HF_TOKEN: ${{ secrets.HF_TOKEN_CI }}
  GGML_NLOOP: 3
  GGML_N_THREADS: 1
  LLAMA_ARG_LOG_COLORS: 1
  LLAMA_ARG_LOG_PREFIX: 1
  LLAMA_ARG_LOG_TIMESTAMPS: 1

jobs:
  gpu-vulkan-nvidia-cm:
    # runs-on: "hf-jobs-t4-small:ubuntu26_04"
    runs-on: [self-hosted, Linux, NVIDIA]

    steps:
      - name: Clone
        id: checkout
        uses: actions/checkout@v6

      # - name: Install dependencies
      #   run: |
      #     sudo apt update
      #     sudo apt install -y build-essential cmake libxcb-xinput0 libxcb-xinerama0 libxcb-cursor-dev libvulkan-dev glslc spirv-headers vulkan-tools mesa-vulkan-drivers libglvnd0 libgl1 libglx0 libegl1 libgles2 libssl-dev time unzip wget python3 python3-venv python3-pip

      # - name: ccache
      #   uses: ggml-org/ccache-action@v1.2.24
      #   with:
      #     restore: false
      #     save: false

      # - name: ccache-buckets-restore
      #   uses: ./.github/actions/ccache-buckets
      #   with:
      #     key: self-hosted-vulkan-nvidia-cm
      #     folder: llama.cpp
      #     hf_bucket: ggml-org/cache

      - name: Test
        id: ggml-ci
        run: |
          vulkaninfo --summary
          GG_BUILD_VULKAN=1 GGML_VK_DISABLE_COOPMAT2=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp

      # - name: ccache-buckets-save
      #   if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
      #   uses: ./.github/actions/ccache-buckets
      #   env:
      #     HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
      #   with:
      #     key: self-hosted-vulkan-nvidia-cm
      #     folder: llama.cpp
      #     evict-old-files: 1d
      #     hf_bucket: ggml-org/cache
      #     save: true

  gpu-vulkan-nvidia-cm2:
    # runs-on: "hf-jobs-t4-small:ubuntu26_04"
    runs-on: [self-hosted, Linux, NVIDIA, COOPMAT2]

    steps:
      - name: Clone
        id: checkout
        uses: actions/checkout@v6

      # - name: Install dependencies
      #   run: |
      #     sudo apt update
      #     sudo apt install -y build-essential cmake libxcb-xinput0 libxcb-xinerama0 libxcb-cursor-dev libvulkan-dev glslc spirv-headers vulkan-tools mesa-vulkan-drivers libglvnd0 libgl1 libglx0 libegl1 libgles2 libssl-dev time unzip wget python3 python3-venv python3-pip

      # - name: ccache
      #   uses: ggml-org/ccache-action@v1.2.24
      #   with:
      #     restore: false
      #     save: false

      # - name: ccache-buckets-restore
      #   uses: ./.github/actions/ccache-buckets
      #   with:
      #     key: self-hosted-vulkan-nvidia-cm2
      #     folder: llama.cpp
      #     hf_bucket: ggml-org/cache

      - name: Test
        id: ggml-ci
        run: |
          vulkaninfo --summary
          GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp

      # - name: ccache-buckets-save
      #   if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
      #   uses: ./.github/actions/ccache-buckets
      #   env:
      #     HF_TOKEN: ${{ secrets.HF_TOKEN_CACHE_OUTPUT }}
      #   with:
      #     key: self-hosted-vulkan-nvidia-cm2
      #     folder: llama.cpp
      #     evict-old-files: 1d
      #     hf_bucket: ggml-org/cache
      #     save: true

  gpu-vulkan-apple:
    runs-on: [self-hosted, macOS, ARM64]

    steps:
      - name: Clone
        id: checkout
        uses: actions/checkout@v6

      - name: Test
        id: ggml-ci
        run: |
          vulkaninfo --summary
          GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp

  gpu-vulkan-intel-linux:
    runs-on: [self-hosted, Linux, Intel]

    steps:
      - name: Clone
        id: checkout
        uses: actions/checkout@v6
        with:
          persist-credentials: false

      - name: Test
        id: ggml-ci
        run: |
          vulkaninfo --summary
          GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp

  gpu-vulkan-intel-windows:
    runs-on: [self-hosted, Windows, X64, Intel]

    steps:
      - name: Clone
        id: checkout
        uses: actions/checkout@v6

      - name: Test
        id: ggml-ci
        shell: C:\msys64\usr\bin\bash.exe --noprofile --norc -eo pipefail "{0}"
        env:
          MSYSTEM: UCRT64
          CHERE_INVOKING: 1
          PATH: C:\msys64\ucrt64\bin;C:\msys64\usr\bin;C:\Windows\System32;${{ env.PATH }}
        run: |
          vulkaninfo --summary
          # Skip python related tests with GG_BUILD_LOW_PERF=1 since Windows MSYS2 UCRT64 currently fails to create
          # a valid python environment for testing
          LLAMA_FATAL_WARNINGS=OFF GG_BUILD_NINJA=1 GG_BUILD_VULKAN=1 GG_BUILD_LOW_PERF=1 ./ci/run.sh ./results/llama.cpp ./mnt/llama.cpp

  # TODO: provision AMD GPU machine
  # amd-vulkan:
  #   runs-on: [self-hosted, Linux, AMD]

  #   steps:
  #     - name: Clone
  #       id: checkout
  #       uses: actions/checkout@v6

  #     - name: Test
  #       id: ggml-ci
  #       run: |
  #         vulkaninfo --summary
  #         GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp