build.yml

# Fast build + unit gate on GitHub-hosted runners. The heavier cross-platform
# LIBERO regression sweep lives in vla-ci.yml (self-hosted).
name: build

on:
  push:
    branches: [main, dev]
  pull_request:
    branches: [main]

concurrency:
  group: build-${{ github.ref }}
  cancel-in-progress: true

jobs:
  py-tooling:
    runs-on: ubuntu-24.04
    steps:
      - uses: actions/checkout@v7
      - uses: actions/setup-python@v7
        with:
          python-version: '3.11'
      - name: converter remap
        run: python tests/py/test_converters.py
      - name: binding struct parity
        run: python tests/py/test_bindings.py

  build-gate:
    runs-on: ubuntu-24.04
    steps:
      - uses: actions/checkout@v7
      - name: deps
        run: |
          sudo apt-get update -qq
          sudo apt-get install -y -qq --no-install-recommends \
            build-essential cmake git ca-certificates pkg-config \
            libzmq3-dev cppzmq-dev libprotobuf-dev protobuf-compiler
      # Read the pin rather than repeat it: two copies drift and the stale one
      # silently reuses the wrong _deps.
      - name: read llama.cpp pin
        id: pin
        run: echo "tag=$(bash scripts/llama_tag.sh)" >> "$GITHUB_OUTPUT"
      # The cached _deps holds compiled ggml-cpu objects, and hosted runners do not
      # all have the same CPU. Jobs that run ctest build with GGML_NATIVE=OFF so a
      # tree compiled on one runner does not hit SIGILL on another (AVX-512).
      - uses: actions/cache@v6
        with:
          path: build/_deps
          key: llama-${{ steps.pin.outputs.tag }}-${{ runner.os }}-asan-generic
      # Everything, not a target list: ctest registers tests this job must build,
      # and a named list goes stale the next time one is added.
      - name: build + ctest (CPU, ASan+UBSan, -Werror)
        run: |
          san="-fsanitize=address,undefined -fno-sanitize-recover=undefined -fno-omit-frame-pointer"
          cmake -B build -DCMAKE_BUILD_TYPE=RelWithDebInfo -DGGML_CUDA=OFF -DGGML_NATIVE=OFF -DVLA_BUILD_TESTS=ON \
            -DVLA_WERROR=ON -DCMAKE_C_FLAGS="$san" -DCMAKE_CXX_FLAGS="$san"
          cmake --build build -j"$(nproc)"
          ctest --test-dir build --output-on-failure
      # Neither patch script runs on a CPU build, and build-cuda patches only on a
      # cache miss, so their anchors would otherwise rot unnoticed until someone
      # configures a fresh CUDA or OpenVINO tree. Patch a copy: the real one is
      # cached.
      - name: patch anchors still apply
        run: |
          cp -r build/_deps/llama-src /tmp/llama-patchtest
          python3 scripts/patch_ggml_cuda_ext_hook.py /tmp/llama-patchtest
          python3 scripts/patch_ggml_sycl_ext_hook.py /tmp/llama-patchtest
          python3 scripts/patch_ggml_sycl_ext_hook.py /tmp/llama-patchtest   # idempotent
          python3 scripts/patch_ggml_openvino.py /tmp/llama-patchtest
          python3 scripts/patch_ggml_openvino.py /tmp/llama-patchtest   # idempotent

  build-cuda:
    runs-on: ubuntu-24.04
    container: nvidia/cuda:12.9.1-devel-ubuntu24.04
    steps:
      - name: deps
        run: |
          apt-get update -qq
          DEBIAN_FRONTEND=noninteractive apt-get install -y -qq --no-install-recommends \
            build-essential cmake git ca-certificates pkg-config python3 \
            libzmq3-dev cppzmq-dev libprotobuf-dev protobuf-compiler
      - uses: actions/checkout@v7
      - name: read llama.cpp pin
        id: pin
        run: echo "tag=$(bash scripts/llama_tag.sh)" >> "$GITHUB_OUTPUT"
      - uses: actions/cache@v6
        with:
          path: build/_deps
          key: llama-${{ steps.pin.outputs.tag }}-${{ runner.os }}-cuda-${{ hashFiles('scripts/patch_ggml_cuda_ext_hook.py') }}
      - name: build (compile only, the runner has no GPU)
        run: |
          cmake -B build -DCMAKE_BUILD_TYPE=Release -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=89-real \
            -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined -DVLA_BUILD_TESTS=ON -DVLA_WERROR=ON
          cmake --build build -j"$(nproc)"

  # Intel GPUs through oneAPI SYCL. The runner has no GPU, so ggml-sycl itself
  # cannot start; ctest still runs everything, and test_foldquant_sycl_op drives
  # the FoldQuant kernels straight through the extension hook on the OpenCL CPU
  # device that the DPC++ compiler package ships, against the CPU reference.
  build-sycl:
    runs-on: ubuntu-24.04
    steps:
      - name: free disk (oneAPI + a SYCL build need ~15 GB)
        run: sudo rm -rf /usr/share/dotnet /usr/local/lib/android /opt/ghc /opt/hostedtoolcache/CodeQL
      - uses: actions/checkout@v7
      - name: deps
        run: |
          wget -qO- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB \
            | sudo gpg --dearmor -o /usr/share/keyrings/oneapi-archive-keyring.gpg
          echo "deb [signed-by=/usr/share/keyrings/oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main" \
            | sudo tee /etc/apt/sources.list.d/oneAPI.list
          sudo apt-get update -qq
          sudo apt-get install -y -qq --no-install-recommends \
            build-essential cmake git ca-certificates pkg-config python3 \
            libzmq3-dev cppzmq-dev libprotobuf-dev protobuf-compiler \
            intel-oneapi-compiler-dpcpp-cpp-2026.1 intel-oneapi-mkl-devel-2026.1 intel-oneapi-dnnl-devel-2026.0
      - name: read llama.cpp pin
        id: pin
        run: echo "tag=$(bash scripts/llama_tag.sh)" >> "$GITHUB_OUTPUT"
      - uses: actions/cache@v6
        with:
          path: build/_deps
          key: llama-${{ steps.pin.outputs.tag }}-${{ runner.os }}-sycl-generic-${{ hashFiles('scripts/patch_ggml_sycl_ext_hook.py') }}
      - name: build + ctest (OpenCL CPU device)
        shell: bash
        env:
          ONEAPI_DEVICE_SELECTOR: opencl:cpu
          VLA_FQ_TEST_REQUIRE_DEVICE: 1
        run: |
          source /opt/intel/oneapi/setvars.sh
          sycl-ls
          cmake -B build -DCMAKE_BUILD_TYPE=Release -DGGML_SYCL=ON -DGGML_NATIVE=OFF \
            -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DVLA_BUILD_TESTS=ON -DVLA_WERROR=ON
          cmake --build build -j"$(nproc)"
          ctest --test-dir build --output-on-failure

  # Intel CPUs, GPUs and NPUs through OpenVINO. The CPU plugin runs on the
  # runner, so test_foldquant_ov_op executes the translated FoldQuant graph and
  # checks it against the CPU reference.
  build-openvino:
    runs-on: ubuntu-24.04
    env:
      OV_ARCHIVE: openvino_toolkit_ubuntu24_2026.4.0.22959.99c81491cc3_x86_64
    steps:
      - uses: actions/checkout@v7
      - name: deps
        run: |
          sudo apt-get update -qq
          sudo apt-get install -y -qq --no-install-recommends \
            build-essential cmake git ca-certificates pkg-config python3 \
            libzmq3-dev cppzmq-dev libprotobuf-dev protobuf-compiler \
            opencl-clhpp-headers ocl-icd-opencl-dev opencl-headers
      - uses: actions/cache@v6
        id: ov
        with:
          path: /opt/openvino
          key: ${{ env.OV_ARCHIVE }}
      - name: OpenVINO runtime
        if: steps.ov.outputs.cache-hit != 'true'
        run: |
          curl -fsSL "https://storage.openvinotoolkit.org/repositories/openvino/packages/2026.4/linux/${OV_ARCHIVE}.tgz" \
            | tar xz -C /tmp
          sudo mv "/tmp/${OV_ARCHIVE}" /opt/openvino
      - name: read llama.cpp pin
        id: pin
        run: echo "tag=$(bash scripts/llama_tag.sh)" >> "$GITHUB_OUTPUT"
      - uses: actions/cache@v6
        with:
          path: build/_deps
          key: llama-${{ steps.pin.outputs.tag }}-${{ runner.os }}-openvino-generic-${{ hashFiles('scripts/patch_ggml_openvino.py') }}
      - name: build + ctest (CPU plugin)
        shell: bash
        run: |
          source /opt/openvino/setupvars.sh
          cmake -B build -DCMAKE_BUILD_TYPE=Release -DGGML_OPENVINO=ON -DGGML_NATIVE=OFF -DVLA_BUILD_TESTS=ON -DVLA_WERROR=ON
          cmake --build build -j"$(nproc)"
          ctest --test-dir build --output-on-failure

  build-backends:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        include:
          - { name: opencl, os: ubuntu-24.04, cmake: -DGGML_OPENCL=ON -DVLA_WERROR=ON }
          - { name: metal,  os: macos-15,     cmake: -DGGML_METAL=ON, test: true }
    steps:
      - uses: actions/checkout@v7
      - name: deps
        if: runner.os == 'Linux'
        run: |
          sudo apt-get update -qq
          sudo apt-get install -y -qq --no-install-recommends \
            build-essential cmake git ca-certificates pkg-config \
            libzmq3-dev cppzmq-dev libprotobuf-dev protobuf-compiler \
            ocl-icd-opencl-dev opencl-headers
      - name: deps
        if: runner.os == 'macOS'
        run: brew install cmake zeromq cppzmq protobuf
      - name: read llama.cpp pin
        id: pin
        run: echo "tag=$(bash scripts/llama_tag.sh)" >> "$GITHUB_OUTPUT"
      - uses: actions/cache@v6
        with:
          path: build/_deps
          key: llama-${{ steps.pin.outputs.tag }}-${{ runner.os }}-${{ matrix.name }}
      - name: build
        run: |
          cmake -B build -DCMAKE_BUILD_TYPE=Release ${{ matrix.cmake }} -DVLA_BUILD_TESTS=ON
          cmake --build build -j"$(getconf _NPROCESSORS_ONLN)"
      - name: ctest
        if: matrix.test
        run: ctest --test-dir build --output-on-failure