cutile-python/.github/workflows/ci.yml at 31eba44120401fb27c4d8186b795ce4373df3a2e · NVIDIA/cutile-python · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
name: CI

on:
  push:
    branches:
      - main
      - "pull-request/[0-9]+"
  workflow_dispatch:

jobs:
  test:
    name: Test ${{ matrix.name }}
    runs-on: ${{ matrix.runner }}
    timeout-minutes: ${{ matrix.timeout }}
    if: ${{ matrix.enabled }}
    strategy:
      fail-fast: false
      matrix:
        include:
          - name: "GPU Linux x86_64"
            runner: linux-amd64-gpu-rtxpro6000-latest-1
            os: linux
            gpu: true
            timeout: 30
            enabled: true
          - name: "CPU Linux x86_64"
            runner: linux-amd64-gpu-rtxpro6000-latest-1
            os: linux
            gpu: false
            timeout: 15
            enabled: true
          - name: "CPU Linux aarch64"
            runner: ubuntu-latest  # TODO: Replace with ARM64 runner
            os: linux
            gpu: false
            timeout: 15
            enabled: false
          - name: "CPU Windows x86_64"
            runner: windows-amd64-gpu-rtxpro6000-latest-1
            os: windows
            gpu: false
            timeout: 15
            enabled: false
          - name: "GPU Windows x86_64"
            runner: windows-amd64-gpu-rtxpro6000-latest-1
            os: windows
            gpu: true
            timeout: 30
            enabled: false

    steps:
    - name: Checkout
      uses: actions/checkout@v4
      timeout-minutes: 5

    - name: Test Docker
      timeout-minutes: 5
      shell: ${{ matrix.os == 'windows' && 'pwsh' || 'bash' }}
      run: |
        docker --version
        docker run --rm hello-world

    - name: Check Hardware
      timeout-minutes: 5
      shell: ${{ matrix.os == 'windows' && 'pwsh' || 'bash' }}
      run: |
        if [ "${{ matrix.os }}" = "windows" ]; then
          Write-Host "=== Host Hardware Information ==="
          if [ "${{ matrix.gpu }}" = "true" ]; then
            nvidia-smi
          else
            Write-Host "CPU-only mode (no GPU access)"
          fi
        else
          echo "=== Host Hardware Information ==="
          lspci | grep -i nvidia || echo "No NVIDIA GPU found via lspci"
          if [ "${{ matrix.gpu }}" = "true" ]; then
            nvidia-smi || echo "nvidia-smi not available on host"
          else
            echo "CPU-only mode (same hardware, no --gpus flag)"
          fi
          if [ "${{ matrix.runner }}" = "ubuntu-latest" ]; then
            echo "Architecture: $(uname -m)"
          fi
        fi

    - name: Test CUDA Image
      timeout-minutes: 10
      shell: ${{ matrix.os == 'windows' && 'pwsh' || 'bash' }}
      run: |
        docker pull nvidia/cuda:13.0.2-base-ubuntu24.04

        if [ "${{ matrix.gpu }}" = "true" ]; then
          echo "=== Testing with GPU access ==="
          docker run --rm --gpus all nvidia/cuda:13.0.2-base-ubuntu24.04 bash -c "
            echo 'Container with GPU access:'
            nvidia-smi || echo 'nvidia-smi not available in container'
            cat /etc/os-release
            echo 'CUDA Runtime version:'
            cat /usr/local/cuda/version.txt 2>/dev/null || echo 'CUDA version file not found'
          " || echo "GPU access failed - may not be available on this runner"
        else
          echo "=== Testing without GPU access (CPU only) ==="
          docker run --rm nvidia/cuda:13.0.2-base-ubuntu24.04 bash -c "
            echo 'CUDA container test successful (CPU only - no GPU access)'
            cat /etc/os-release
            which nvcc || echo 'nvcc not in PATH'
            nvidia-smi 2>/dev/null || echo 'nvidia-smi not available without --gpus flag (expected)'
          "
        fi