llamacpp.yaml

name: Build llama.cpp in CANN container

on:
  workflow_dispatch:
  pull_request:
    paths:
      - '.github/workflows/llamacpp.yaml'
      - 'requirements/**'
  schedule:
    - cron: "0 0 * * *"

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

# Bash shells do not use ~/.profile or ~/.bashrc so these shells need to be explicitly
# declared as "shell: bash -el {0}" on steps that need to be properly activated.
# It's used to activate ascend-toolkit environment variables.
defaults:
  run:
    shell: bash -el {0}

jobs:
  openeuler-arm64-test:
    if: ${{ github.event_name != 'pull_request' || contains(github.event.pull_request.labels.*.name, 'Ascend NPU') }}
    name: Build llama.cpp on OpenEuler for Arm64
    runs-on: ubuntu-24.04-arm
    strategy:
      matrix:
        cann:
          - '8.0.rc3.beta1-910b-openeuler22.03-py3.10'
        device:
          - 'ascend910b3'
        build:
          - 'Release'
    container: ascendai/cann:${{ matrix.cann }}
    steps:
      - name: Install dependencies
        run: |
          yum update -y
          yum install -y git gcc gcc-c++ make cmake

      - name: Checkout
        uses: actions/checkout@v4

      - name: Checkout llama.cpp
        uses: actions/checkout@v4
        with:
          repository: ggerganov/llama.cpp
          path: llama.cpp

      - name: Build llama.cpp
        working-directory: llama.cpp
        run: |
          mkdir build
          cd build

          export LD_LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/$(uname -m)-linux/devlib/:${LD_LIBRARY_PATH}

          cmake .. -DCMAKE_BUILD_TYPE=${{ matrix.build }} -DGGML_CANN=on -DSOC_TYPE=${{ matrix.device }}
          cmake --build . -j $(nproc)