spack/share/spack/gitlab/cloud_pipelines/stacks/ml-linux-x86_64-rocm/spack.yaml

spack:
  view: false

  concretizer:
    reuse: false
    unify: false

  config:
    build_jobs: 32
    concretizer: clingo
    install_tree:
      root: /home/software/spack
      padded_length: 384
      projections:
        all: "{architecture}/{compiler.name}-{compiler.version}/{name}-{version}-{hash}"

  packages:
    all:
      target: [x86_64_v3]
      variants: ~cuda+rocm amdgpu_target=gfx90a
    gl:
      require: "osmesa"
    py-torch:
      # Does not yet support Spack-installed ROCm
      require: ~rocm

  definitions:
    - packages:
      # Horovod
      - py-horovod

      # Hugging Face
      - py-transformers

      # JAX
      - py-jax
      - py-jaxlib

      # Keras
      - py-keras
      - py-keras-applications
      - py-keras-preprocessing
      - py-keras2onnx

      # PyTorch
      # Does not yet support Spack-install ROCm
      # - py-botorch
      # - py-efficientnet-pytorch
      # - py-gpytorch
      # - py-kornia
      # - py-pytorch-gradual-warmup-lr
      # - py-pytorch-lightning
      # - py-segmentation-models-pytorch
      # - py-timm
      # - py-torch
      # - py-torch-cluster
      # - py-torch-geometric
      # - py-torch-nvidia-apex
      # - py-torch-scatter
      # - py-torch-sparse
      # - py-torch-spline-conv
      # - py-torchaudio
      # - py-torchdata
      # - py-torchfile
      # - py-torchgeo
      # - py-torchmeta
      # - py-torchmetrics
      # - py-torchtext
      # - py-torchvision
      # - py-vector-quantize-pytorch

      # scikit-learn
      - py-scikit-learn
      - py-scikit-learn-extra

      # TensorBoard
      - py-tensorboard
      - py-tensorboard-data-server
      - py-tensorboard-plugin-wit
      - py-tensorboardx

      # TensorFlow
      - py-tensorflow
      - py-tensorflow-datasets
      - py-tensorflow-estimator
      - py-tensorflow-hub
      - py-tensorflow-metadata
      - py-tensorflow-probability

      # XGBoost
      - py-xgboost
      # - r-xgboost
      - xgboost

    - arch:
      - target=x86_64_v3

  specs:
    - matrix:
      - [$packages]
      - [$arch]

  mirrors: { "mirror": "s3://spack-binaries/develop/ml-linux-x86_64-rocm" }

  gitlab-ci:
    script:
      - uname -a || true
      - grep -E 'vendor|model name' /proc/cpuinfo 2>/dev/null | sort -u || head -n10 /proc/cpuinfo 2>/dev/null || true
      - nproc
      - curl -Lfs 'https://github.com/JuliaBinaryWrappers/GNUMake_jll.jl/releases/download/GNUMake-v4.3.0+1/GNUMake.v4.3.0.x86_64-linux-gnu.tar.gz' -o gmake.tar.gz
      - printf 'fef1f59e56d2d11e6d700ba22d3444b6e583c663d6883fd0a4f63ab8bd280f0f gmake.tar.gz' | sha256sum --check --strict --quiet
      - tar -xzf gmake.tar.gz -C /usr bin/make 2> /dev/null
      - . "./share/spack/setup-env.sh"
      - spack --version
      - spack arch
      - cd ${SPACK_CONCRETE_ENV_DIR}
      - spack env activate --without-view .
      - spack config add "config:install_tree:projections:${SPACK_JOB_SPEC_PKG_NAME}:'morepadding/{architecture}/{compiler.name}-{compiler.version}/{name}-{version}-{hash}'"
      - mkdir -p ${SPACK_ARTIFACTS_ROOT}/user_data
      - if [[ -r /mnt/key/intermediate_ci_signing_key.gpg ]]; then spack gpg trust /mnt/key/intermediate_ci_signing_key.gpg; fi
      - if [[ -r /mnt/key/spack_public_key.gpg ]]; then spack gpg trust /mnt/key/spack_public_key.gpg; fi
      - spack --color=always --backtrace ci rebuild > >(tee ${SPACK_ARTIFACTS_ROOT}/user_data/pipeline_out.txt) 2> >(tee ${SPACK_ARTIFACTS_ROOT}/user_data/pipeline_err.txt >&2)
    after_script:
      - cat /proc/loadavg || true

    match_behavior: first
    mappings:
      - match:
          - llvm
          - llvm-amdgpu
          - py-tensorflow
          - py-torch
          - rocblas
        runner-attributes:
          tags: [ "spack", "huge", "x86_64_v4" ]
          variables:
            CI_JOB_SIZE: huge
            KUBERNETES_CPU_REQUEST: 11000m
            KUBERNETES_MEMORY_REQUEST: 42G
      - match:
          - "@:"
        runner-attributes:
          tags: [ "spack", "large", "x86_64_v4" ]
          variables:
            CI_JOB_SIZE: large
            KUBERNETES_CPU_REQUEST: 8000m
            KUBERNETES_MEMORY_REQUEST: 12G

    image: { "name": "ghcr.io/spack/e4s-amazonlinux-2:v2022-03-21", "entrypoint": [""] }

    broken-specs-url: "s3://spack-binaries/broken-specs"

    service-job-attributes:
      before_script:
        - . "./share/spack/setup-env.sh"
        - spack --version
      image: { "name": "ghcr.io/spack/e4s-amazonlinux-2:v2022-03-21", "entrypoint": [""] }
      tags: ["spack", "public", "x86_64_v4"]

    signing-job-attributes:
      image: { "name": "ghcr.io/spack/notary:latest", "entrypoint": [""] }
      tags: ["spack", "aws"]
      script:
        - aws s3 sync --exclude "*" --include "*spec.json*" ${SPACK_REMOTE_MIRROR_OVERRIDE}/build_cache /tmp
        - /sign.sh
        - aws s3 sync --exclude "*" --include "*spec.json.sig*" /tmp ${SPACK_REMOTE_MIRROR_OVERRIDE}/build_cache

  cdash:
    build-group: Machine Learning
    url: https://cdash.spack.io
    project: Spack Testing
    site: Cloud Gitlab Infrastructure