Merge branch 'main' into py313

.github/actions/setup/action.yml

@@ -47,7 +47,22 @@ runs:
       run: |
         env
 
+    - name: Set up CTK cache variable
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: |
+        echo "CTK_CACHE_KEY=mini-ctk-${{ inputs.cuda-version }}-${{ inputs.host-platform }}" >> $GITHUB_ENV
+        echo "CTK_CACHE_FILENAME=mini-ctk-${{ inputs.cuda-version }}-${{ inputs.host-platform }}.tar.gz" >> $GITHUB_ENV
+
+    - name: Download CTK cache
+      id: ctk-get-cache
+      uses: actions/cache/restore@v4
+      continue-on-error: true
+      with:
+        key: ${{ env.CTK_CACHE_KEY }}
+        path: ./${{ env.CTK_CACHE_FILENAME }}
+
     - name: Get CUDA components
+      if: ${{ steps.ctk-get-cache.outputs.cache-hit != 'true' }}
       shell: bash --noprofile --norc -xeuo pipefail {0}
       run: |
         CUDA_PATH="./cuda_toolkit"
@@ -90,15 +105,43 @@ runs:
         }
 
         # Get headers and shared libraries in place
+        # Note: the existing artifact would need to be manually deleted (ex: through web UI)
+        # if this list is changed, as the artifact actions do not offer any option for us to
+        # invalidate the artifact.
         populate_cuda_path cuda_nvcc
         populate_cuda_path cuda_cudart
         populate_cuda_path cuda_nvrtc
         populate_cuda_path cuda_profiler_api
+        populate_cuda_path libnvjitlink
         ls -l $CUDA_PATH
 
+        # Prepare the cache
+        # Note: try to escape | and > ...
+        tar -czvf ${CTK_CACHE_FILENAME} ${CUDA_PATH}
+
         # Note: the headers will be copied into the cibuildwheel manylinux container,
         # so setting the CUDA_PATH env var here is meaningless.
 
+    - name: Upload CTK cache
+      if: ${{ always() &&
+              steps.ctk-get-cache.outputs.cache-hit != 'true' }}
+      uses: actions/cache/save@v4
+      with:
+        key: ${{ env.CTK_CACHE_KEY }}
+        path: ./${{ env.CTK_CACHE_FILENAME }}
+
+    - name: Restore CTK cache
+      if: ${{ steps.ctk-get-cache.outputs.cache-hit == 'true' }}
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: |
+        ls -l
+        CUDA_PATH="./cuda_toolkit"
+        tar -xzvf $CTK_CACHE_FILENAME
+        ls -l $CUDA_PATH
+        if [ ! -d "$CUDA_PATH/include" ]; then
+          exit 1
+        fi
+
     - name: Set environment variables
       shell: bash --noprofile --norc -xeuo pipefail {0}
       run: |

.github/actions/test/action.yml

@@ -0,0 +1,93 @@
+name: test
+
+description: Run tests in specified project
+
+inputs:
+  test-options:
+    required: true
+    type: string
+
+runs:
+  using: composite
+  steps:
+    - name: Run nvidia-smi to make sure GPU is working
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: nvidia-smi
+
+    - name: Download bindings build artifacts
+      uses: actions/download-artifact@v4
+      with:
+        name: ${{ env.CUDA_BINDINGS_ARTIFACT_NAME }}
+        path: ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }}
+
+    - name: Display structure of downloaded bindings artifacts
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: |
+        pwd
+        ls -lahR $CUDA_BINDINGS_ARTIFACTS_DIR
+
+    - name: Download core build artifacts
+      uses: actions/download-artifact@v4
+      with:
+        name: ${{ env.CUDA_CORE_ARTIFACT_NAME }}
+        path: ${{ env.CUDA_CORE_ARTIFACTS_DIR }}
+
+    - name: Display structure of downloaded core build artifacts
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: |
+        pwd
+        ls -lahR $CUDA_CORE_ARTIFACTS_DIR
+
+    - name: Set up Python ${{ env.PYTHON_VERSION }}
+      uses: actions/setup-python@v5
+      with:
+        python-version: ${{ env.PYTHON_VERSION }}
+
+    - name: Set up CTK cache variable
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: |
+        echo "CTK_CACHE_KEY=mini-ctk-${CTK_BUILD_VER}-${HOST_PLATFORM}" >> $GITHUB_ENV
+        echo "CTK_CACHE_FILENAME=mini-ctk-${CTK_BUILD_VER}-${HOST_PLATFORM}.tar.gz" >> $GITHUB_ENV
+
+    - name: Download CTK cache
+      id: ctk-get-cache
+      uses: actions/cache/restore@v4
+      continue-on-error: true
+      with:
+        key: ${{ env.CTK_CACHE_KEY }}
+        path: ./${{ env.CTK_CACHE_FILENAME }}
+        fail-on-cache-miss: true
+
+    - name: Restore CTK cache
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: |
+        ls -l
+        CUDA_PATH="$(pwd)/cuda_toolkit"
+        tar -xzvf $CTK_CACHE_FILENAME
+        ls -l $CUDA_PATH
+        if [ ! -d "$CUDA_PATH/include" ]; then
+          exit 1
+        fi
+
+        echo "CUDA_PATH=$CUDA_PATH" >> $GITHUB_ENV
+        echo "PATH=$PATH:$CUDA_PATH/bin" >> $GITHUB_ENV
+        echo "LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$CUDA_PATH/lib" >> $GITHUB_ENV
+
+    - name: Run test / analysis
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: |
+        REPO_DIR=$(pwd)
+
+        cd "${CUDA_BINDINGS_ARTIFACTS_DIR}"
+        pip install *.whl
+
+        cd "${CUDA_CORE_ARTIFACTS_DIR}"
+        pip install *.whl
+
+        cd "${REPO_DIR}/cuda_bindings"
+        pip install -r requirements.txt
+        pytest tests/
+        #pytest tests/cython
+
+        cd "${REPO_DIR}/cuda_core"
+        pytest -rxXs tests/

.github/workflows/ci-gh.yml

@@ -1,4 +1,4 @@
-name: Build and test
+name: "CI"
 
 concurrency:
   group: ${{ startsWith(github.ref_name, 'main') && format('unique-{0}', github.run_id) || format('ci-build-and-test-on-{0}-from-{1}', github.event_name, github.ref_name) }}
@@ -11,8 +11,7 @@ on:
       - "main"
 
 jobs:
-  build-and-test:
-    name: Build and test (${{ matrix.host-platform }}, ${{ matrix.target-device }}, ${{ matrix.build-mode }})
+  ci:
     strategy:
       fail-fast: false
       matrix:
@@ -36,6 +35,7 @@ jobs:
           # Note: this is for build-time only; the test-time matrix needs to be
           # defined separately.
           - "12.6.2"
+    name: "CI"
     uses:
       ./.github/workflows/gh-build-and-test.yml
     with:

.github/workflows/gh-build-and-test.yml

@@ -25,21 +25,100 @@ on:
 
 jobs:
   build:
+    name: Build (${{ inputs.host-platform }}, Python "${{ inputs.python-version }}")
     if: ${{ github.repository_owner == 'nvidia' }}
-    uses:
-      ./.github/workflows/gh-build.yml
-    with:
-      client-repo: ${{ github.event.repository.name }}
-      target-device: ${{ inputs.target-device }}
-      runs-on: ${{ (inputs.host-platform == 'linux-x64' && 'linux-amd64-cpu8') ||
-                   (inputs.host-platform == 'linux-aarch64' && 'linux-arm64-cpu8') ||
-                   (inputs.host-platform == 'win-x64' && 'windows-2019') }}
-                 #  (inputs.host-platform == 'win-x64' && 'windows-amd64-cpu8') }}
-      build-type: ${{ inputs.build-type }}
-      host-platform: ${{ inputs.host-platform }}
-      build-mode: ${{ inputs.build-mode }}
-      upload-enabled: ${{ inputs.upload-enabled }}
-      python-version: ${{ inputs.python-version }}
-      cuda-version: ${{ inputs.cuda-version }}
-      dependencies-file: ""
-    secrets: inherit
+    permissions:
+      id-token: write # This is required for configure-aws-credentials
+      contents: read  # This is required for actions/checkout
+    runs-on: ${{ (inputs.host-platform == 'linux-x64' && 'linux-amd64-cpu8') ||
+                 (inputs.host-platform == 'linux-aarch64' && 'linux-arm64-cpu8') ||
+                 (inputs.host-platform == 'win-x64' && 'windows-2019') }}
+               #  (inputs.host-platform == 'win-x64' && 'windows-amd64-cpu8') }}
+    outputs:
+      CUDA_CORE_ARTIFACT_NAME: ${{ steps.pass_env.outputs.CUDA_CORE_ARTIFACT_NAME }}
+      CUDA_CORE_ARTIFACTS_DIR: ${{ steps.pass_env.outputs.CUDA_CORE_ARTIFACTS_DIR }}
+      CUDA_BINDINGS_ARTIFACT_NAME: ${{ steps.pass_env.outputs.CUDA_BINDINGS_ARTIFACT_NAME }}
+      CUDA_BINDINGS_ARTIFACTS_DIR: ${{ steps.pass_env.outputs.CUDA_BINDINGS_ARTIFACTS_DIR }}
+    steps:
+      - name: Checkout ${{ github.event.repository.name }}
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Set up build environment
+        uses: ./.github/actions/setup
+        with:
+          client-repo: ${{ github.event.repository.name }}
+          build-type: ${{ inputs.build-type }}
+          target-device: "${{ inputs.target-device }}"
+          host-platform: ${{ inputs.host-platform }}
+          build-mode: ${{ inputs.build-mode }}
+          upload-enabled: ${{ inputs.upload-enabled }}
+          python-version: ${{ inputs.python-version }}
+          cuda-version: ${{ inputs.cuda-version }}
+
+      - name: Call build action
+        uses: ./.github/actions/build
+        with:
+          build-type: ${{ inputs.build-type }}
+          target-device: "${{ inputs.target-device }}"
+          host-platform: ${{ inputs.host-platform }}
+          upload-enabled: ${{ inputs.upload-enabled }}
+
+      - name: Pass environment variables
+        id: pass_env
+        run: |
+          echo "CUDA_CORE_ARTIFACT_NAME=${CUDA_CORE_ARTIFACT_NAME}" >> $GITHUB_OUTPUT
+          echo "CUDA_CORE_ARTIFACTS_DIR=${CUDA_CORE_ARTIFACTS_DIR}" >> $GITHUB_OUTPUT
+          echo "CUDA_BINDINGS_ARTIFACT_NAME=${CUDA_BINDINGS_ARTIFACT_NAME}" >> $GITHUB_OUTPUT
+          echo "CUDA_BINDINGS_ARTIFACTS_DIR=${CUDA_BINDINGS_ARTIFACTS_DIR}" >> $GITHUB_OUTPUT
+
+  test:
+    # TODO: improve the name once a separate test matrix is defined
+    name: Test (CUDA ${{ inputs.cuda-version }})
+    # TODO: enable testing once linux-aarch64 & win-64 GPU runners are up
+    if: ${{ (github.repository_owner == 'nvidia') &&
+             startsWith(inputs.host-platform, 'linux-x64') }}
+    permissions:
+      id-token: write # This is required for configure-aws-credentials
+      contents: read  # This is required for actions/checkout
+    runs-on: ${{ (inputs.host-platform == 'linux-x64' && 'linux-amd64-gpu-v100-latest-1') }}
+    # TODO: use a different (nvidia?) container, or just run on bare image
+    container:
+      options: -u root --security-opt seccomp=unconfined --privileged --shm-size 16g
+      image: condaforge/miniforge3:latest
+      env:
+        NVIDIA_VISIBLE_DEVICES: ${{ env.NVIDIA_VISIBLE_DEVICES }}
+    needs:
+      - build
+    steps:
+      - name: Checkout ${{ github.event.repository.name }}
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      # TODO: we probably don't need this?
+      # - name: Setup
+      #   if: ${{ inputs.has-built != 'true' }}
+      #   uses: ./.github/actions/setup
+      #   with:
+      #     client-repo: ${{ github.event.repository.name }}
+      #     build-type: ${{ inputs.build-type }}
+      #     target-device: "${{ inputs.target-device }}"
+      #     host-platform: ${{ inputs.host-platform }}
+      #     build-mode: ${{ inputs.build-mode }}
+      #     upload-enabled: ${{ inputs.upload-enabled }}
+      #     python-version: ${{ inputs.python-version }}
+
+      - name: Call test action
+        uses: ./.github/actions/test
+        with:
+          test-options: ${{ inputs.build-type }}
+        env:
+          CUDA_CORE_ARTIFACT_NAME: ${{ needs.build.outputs.CUDA_CORE_ARTIFACT_NAME }}
+          CUDA_CORE_ARTIFACTS_DIR: ${{ needs.build.outputs.CUDA_CORE_ARTIFACTS_DIR }}
+          CUDA_BINDINGS_ARTIFACT_NAME: ${{ needs.build.outputs.CUDA_BINDINGS_ARTIFACT_NAME }}
+          CUDA_BINDINGS_ARTIFACTS_DIR: ${{ needs.build.outputs.CUDA_BINDINGS_ARTIFACTS_DIR }}
+          PYTHON_VERSION: ${{ inputs.python-version }}
+          CTK_BUILD_VER: ${{ inputs.cuda-version }}
+          HOST_PLATFORM: ${{ inputs.host-platform }}