lightning/.azure/gpu-tests.yml

# Python package
# Create and test a Python package on multiple Python versions.
# Add steps that analyze code, save the dist with the build record, publish to a PyPI-compatible index, and more:
# https://docs.microsoft.com/azure/devops/pipelines/languages/python

trigger:
  tags:
    include:
      - '*'
  branches:
    include:
      - "master"
      - "release/*"
      - "refs/tags/*"

pr:
  - "master"
  - "release/*"

variables:
  - name: continue
    value: '1'

jobs:
  - job: testing
    strategy:
      matrix:
        'PyTorch - stable':
          image: "pytorchlightning/pytorch_lightning:base-cuda-py3.9-torch1.12-cuda11.3.1"
    # how long to run the job before automatically cancelling
    timeoutInMinutes: "80"
    # how much time to give 'run always even if cancelled tasks' before stopping them
    cancelTimeoutInMinutes: "2"
    pool: azure-jirka-spot
    container:
      image: $(image)
      # default shm size is 64m. Increase it to avoid:
      # 'Error while creating shared memory: unhandled system error, NCCL version 2.7.8'
      options: "--runtime=nvidia -e NVIDIA_VISIBLE_DEVICES=all --shm-size=512m"
    workspace:
      clean: all

    steps:

    - bash: |
        CHANGED_FILES=$(git diff --name-status origin/master -- . | awk  '{print $2}')
        FILTER='src/pytorch_lightning|requirements/pytorch|tests/tests_pytorch|examples/pl_*'
        echo $CHANGED_FILES > changed_files.txt
        MATCHES=$(cat changed_files.txt | grep -E $FILTER)
        echo $MATCHES
        if [ -z "$MATCHES" ]; then
            echo "Skip"
            echo "##vso[task.setvariable variable=continue]0"
        else
            echo "Continue"
            echo "##vso[task.setvariable variable=continue]1"
        fi        
      displayName: Skipper

    - bash: |
        lspci | egrep 'VGA|3D'
        whereis nvidia
        nvidia-smi
        which python && which pip
        python --version
        pip --version
        pip list        
      displayName: 'Image info & NVIDIA'
      condition: eq(variables['continue'], '1')

    - bash: |
        set -e
        python -c "fname = 'requirements/pytorch/strategies.txt' ; lines = [line for line in open(fname).readlines() if 'horovod' not in line] ; open(fname, 'w').writelines(lines)"
        python -c "fname = 'requirements/pytorch/strategies.txt' ; lines = [line for line in open(fname).readlines() if 'bagua' not in line] ; open(fname, 'w').writelines(lines)"
        CUDA_VERSION_MM=$(python -c "import torch ; print(''.join(map(str, torch.version.cuda.split('.')[:2])))")
        pip install "bagua-cuda$CUDA_VERSION_MM>=0.9.0"
        pip install -e .[strategies]
        pip install -U deepspeed  # TODO: remove when docker images are upgraded
        pip install --requirement requirements/pytorch/devel.txt
        pip list        
      env:
        PACKAGE_NAME: pytorch
        FREEZE_REQUIREMENTS: 1
      displayName: 'Install dependencies'
      condition: eq(variables['continue'], '1')

    - bash: |
        set -e
        python requirements/collect_env_details.py
        python -c "import torch ; mgpu = torch.cuda.device_count() ; assert mgpu >= 2, f'GPU: {mgpu}'"
        python requirements/pytorch/check-avail-strategies.py
        python requirements/pytorch/check-avail-extras.py        
      displayName: 'Env details'
      condition: eq(variables['continue'], '1')

    - bash: bash .actions/pull_legacy_checkpoints.sh
      displayName: 'Get legacy checkpoints'
      condition: eq(variables['continue'], '1')

    - bash: python -m coverage run --source pytorch_lightning -m pytest
      workingDirectory: src/pytorch_lightning
      displayName: 'Testing: PyTorch doctests'
      condition: eq(variables['continue'], '1')

    - bash: python -m coverage run --source pytorch_lightning -m pytest --ignore benchmarks -v --junitxml=$(Build.StagingDirectory)/test-results.xml --durations=50
      env:
        PL_RUN_CUDA_TESTS: "1"
      workingDirectory: tests/tests_pytorch
      displayName: 'Testing: PyTorch standard'
      timeoutInMinutes: "35"
      condition: eq(variables['continue'], '1')

    - bash: bash run_standalone_tests.sh
      workingDirectory: tests/tests_pytorch
      env:
        PL_USE_MOCKED_MNIST: "1"
        PL_RUN_CUDA_TESTS: "1"
      displayName: 'Testing: PyTorch standalone tests'
      timeoutInMinutes: "35"
      condition: eq(variables['continue'], '1')

    - bash: |
        python -m coverage report
        python -m coverage xml
        python -m coverage html
        python -m codecov --token=$(CODECOV_TOKEN) --commit=$(Build.SourceVersion) --flags=gpu,pytest --name="GPU-coverage" --env=linux,azure
        ls -l        
      workingDirectory: tests/tests_pytorch
      displayName: 'Statistics'
      condition: eq(variables['continue'], '1')

    - task: PublishTestResults@2
      displayName: 'Publish test results'
      inputs:
        testResultsFiles: '$(Build.StagingDirectory)/test-results.xml'
        testRunTitle: '$(Agent.OS) - $(Build.DefinitionName) - Python $(python.version)'
      condition: and(succeededOrFailed(), eq(variables['continue'], '1'))

    - script: |
        set -e
        bash run_ddp_examples.sh
        bash run_pl_examples.sh --trainer.accelerator=gpu --trainer.devices=1
        bash run_pl_examples.sh --trainer.accelerator=gpu --trainer.devices=2 --trainer.strategy=ddp
        bash run_pl_examples.sh --trainer.accelerator=gpu --trainer.devices=2 --trainer.strategy=ddp --trainer.precision=16        
      workingDirectory: examples
      env:
        PL_USE_MOCKED_MNIST: "1"
      displayName: 'Testing: PyTorch examples'
      condition: eq(variables['continue'], '1')

    - bash: python -m pytest benchmarks -v --maxfail=2 --durations=0
      workingDirectory: tests/tests_pytorch
      env:
        PL_RUN_CUDA_TESTS: "1"
      displayName: 'Testing: PyTorch benchmarks'
      condition: eq(variables['continue'], '1')