Merge branch 'main' into 109-benchmarks

Author: leofang

.github/ISSUE_TEMPLATE/release_checklist.yml

@@ -1,39 +1,33 @@
 name: Release Checklist
-description: Release managers need to go through this check list and ensure all tasks are done
+description: Public checklist for release managers making a subpackage release
+labels: ["P0"]
+
 body:
   - type: markdown
     attributes:
       value: |
-        Please go through this checklist and ensure all tasks are completed.
-
-  - type: checkboxes
-    id: cuda-bindings-tasks
-    attributes:
-      label: Tasks for cuda-bindings / cuda-python release
-      options:
-        - label: "Push any internal updates for accommodating a new CTK release to the public (**IMPORTANT**: Need to wait for CTK posting!)"
-        - label: "If it is a major release, create a new branch to backport to and update the branch name [here](../BACKPORT_BRANCH)"
-        - label: Follow the check list for `cuda-core` below for the remaining steps
+        This checklist is for cuda-core releases as well as cuda-bindings patches. Please go through this checklist and ensure all tasks are completed.
 
   - type: checkboxes
-    id: cuda-core-tasks
+    id: subpackage-tasks
     attributes:
-      label: Tasks for cuda-core release
+      label: Tasks for cuda-core and cuda-bindings patch release
       options:
         - label: File an internal nvbug to communicate test plan & release schedule with QA
         - label: Ensure all pending PRs are reviewed, tested, and merged
-        - label: Update the dependency requirements, if needed
+        - label: Check (or update if needed) the dependency requirements
         - label: Bump the version
         - label: Create a public rc tag
         - label: "Point QA to fetch public artifacts (wheels) from the GHA run ID, example: `gh run download 12323257563 -p \"cuda-core*\" -R NVIDIA/cuda-python`"
         - label: Wait for QA reports and fix any issues found
-        - label: Finalize the doc update, including release notes
+        - label: "Finalize the doc update, including release notes (\"Note: Touching docstrings/type annotations in code is OK during code freeze, apply your best judgement!\")"
         - label: Update the docs for the new version
         - label: Create a public release tag
         - label: If any code change happens, rebuild the wheels from the new tag
+        - label: Update the conda recipe & release conda packages
+        - label: Upload conda packages to nvidia channel
         - label: Upload wheels to PyPI
+        - label: Post-release QA
         - label: Finalize the announcement update
         - label: Send out the announcement internally
         - label: Send out the announcement externally (GitHub Release -> Announcement)
-        - label: Update the conda recipe & release conda packages
-        - label: Post-release QA

.github/actions/doc_preview/action.yml

@@ -0,0 +1,65 @@
+name: Preview or clean up docs built from PRs
+
+# A re-implementation based on the logic of https://github.com/rossjrw/pr-preview-action/blob/41a957c44a456a34718e9bcf825363194db5e6d5/README.md, due to limitations illustrated in NVIDIA/cuda-python#380.
+
+inputs:
+  source-folder:
+    required: true
+    type: string
+  pr-number:
+    required: true
+    type: string
+
+runs:
+  using: composite
+  steps:
+    # The steps below are executed only when testing in a PR.
+    # Note: the PR previews will be removed once merged to main (see below) 
+    - name: Deploy doc preview
+      if: ${{ github.ref_name != 'main' }}
+      uses: JamesIves/github-pages-deploy-action@v4
+      with:
+        git-config-name: cuda-python-bot
+        git-config-email: cuda-python-bot@users.noreply.github.com
+        folder: ${{ inputs.source-folder }}
+        target-folder: docs/pr-preview/pr-${{ inputs.pr-number }}/
+        commit-message: "Deploy doc preview for PR ${{ inputs.pr-number }} (${{ github.sha }})"
+    
+    - name: Leave a comment after deployment
+      if: ${{ github.ref_name != 'main' }}
+      uses: marocchino/sticky-pull-request-comment@v2
+      with:
+        header: pr-preview
+        number: ${{ inputs.pr-number }}
+        skip_unchanged: true
+        message: |
+          Doc Preview CI
+          :---:
+          | <p></p> :rocket: View preview at <br> https://nvidia.github.io/cuda-python/pr-preview/pr-${{ inputs.pr-number }}/ <br>
+          | <br> https://nvidia.github.io/cuda-python/pr-preview/pr-${{ inputs.pr-number }}/cuda-core/ <br>
+          | <br> https://nvidia.github.io/cuda-python/pr-preview/pr-${{ inputs.pr-number }}/cuda-bindings/ <br><br>
+          | <h6><br> Preview will be ready when the GitHub Pages deployment is complete. <br><br></h6>
+    
+    # The steps below are executed only when building on main.    
+    - name: Remove doc preview
+      if: ${{ github.ref_name == 'main' }}
+      uses: JamesIves/github-pages-deploy-action@v4
+      with:
+        git-config-name: cuda-python-bot
+        git-config-email: cuda-python-bot@users.noreply.github.com
+        folder: ${{ inputs.source-folder }}
+        target-folder: docs/pr-preview/pr-${{ inputs.pr-number }}/
+        commit-message: "Clean up doc preview for PR ${{ inputs.pr-number }} (${{ github.sha }})"
+    
+    - name: Leave a comment after removal
+      if: ${{ github.ref_name == 'main' }}
+      uses: marocchino/sticky-pull-request-comment@v2
+      with:
+        header: pr-preview
+        number: ${{ inputs.pr-number }}
+        hide_and_recreate: true
+        hide_classify: "OUTDATED"
+        message: |
+          Doc Preview CI
+          :---:
+          Preview removed because the pull request was closed or merged.

.github/actions/fetch_ctk/action.yml

@@ -123,4 +123,4 @@ runs:
         echo "CUDA_PATH=${CUDA_PATH}" >> $GITHUB_ENV
         echo "CUDA_HOME=${CUDA_PATH}" >> $GITHUB_ENV
         echo "${CUDA_PATH}/bin" >> $GITHUB_PATH
-        echo "LD_LIBRARY_PATH=${LD_LIBRARY_PATH:-}:${CUDA_PATH}/lib" >> $GITHUB_ENV
+        echo "LD_LIBRARY_PATH=${LD_LIBRARY_PATH:-}:${CUDA_PATH}/lib:${CUDA_PATH}/nvvm/lib64" >> $GITHUB_ENV

.github/actions/get_pr_number/action.yml

@@ -0,0 +1,54 @@
+name: Get the PR number
+
+description: Get the PR number without relying on the pull_request* event triggers.
+
+runs:
+  using: composite
+  steps:
+    - name: Get PR info (non-main branch)
+      if: ${{ github.ref_name != 'main' }}
+      uses: nv-gha-runners/get-pr-info@main
+      id: get-pr-info
+
+    - name: Extract PR number (non-main branch)
+      if: ${{ github.ref_name != 'main' }}
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: |
+        trap 'echo "Error at line $LINENO"; exit 1' ERR
+        PR_NUMBER="${{ fromJSON(steps.get-pr-info.outputs.pr-info).number }}"
+        if [[ -z "$PR_NUMBER" ]]; then
+          echo "Cannot extract PR number for ref: ${{ github.ref_name }}"
+          exit 1
+        fi
+        echo "PR_NUMBER=$PR_NUMBER" >> $GITHUB_ENV
+        echo "BUILD_PREVIEW=1" >> $GITHUB_ENV
+
+    - name: Get PR data (main branch)
+      if: ${{ github.ref_name == 'main' }}
+      uses: actions/github-script@v7
+      id: get-pr-data
+      with:
+        script: |
+          const prs = await github.rest.repos.listPullRequestsAssociatedWithCommit({
+            commit_sha: context.sha,
+            owner: context.repo.owner,
+            repo: context.repo.repo,
+          });
+          if (!prs.data.length) {
+            core.setFailed("No PR associated with this commit on 'main'.");
+          } else {
+            return prs.data[0];
+          }
+
+    - name: Extract PR number (main branch)
+      if: ${{ github.ref_name == 'main' }}
+      shell: bash --noprofile --norc -xeuo pipefail {0}
+      run: |
+        trap 'echo "Error at line $LINENO"; exit 1' ERR
+        PR_NUMBER="${{ fromJSON(steps.get-pr-data.outputs.result).number }}"
+        if [[ -z "$PR_NUMBER" ]]; then
+          echo "No associated PR found for the commit in 'main'."
+          exit 1
+        fi
+        echo "PR_NUMBER=$PR_NUMBER" >> $GITHUB_ENV
+        echo "BUILD_LATEST=1" >> $GITHUB_ENV

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

@@ -31,7 +31,7 @@ jobs:
           - "3.9"
         cuda-version:
           # Note: this is for build-time only.
-          - "12.6.2"
+          - "12.8.0"
     name: Build (${{ matrix.host-platform }}, Python ${{ matrix.python-version }}, CUDA ${{ matrix.cuda-version }})
     if: ${{ github.repository_owner == 'nvidia' }}
     permissions:
@@ -76,7 +76,7 @@ jobs:
             REPO_DIR=$(cygpath -w $PWD)
           fi
   
-          echo "PARALLEL_LEVEL=$(nproc)" >> $GITHUB_ENV
+          echo "CUDA_BINDINGS_PARALLEL_LEVEL=$(nproc)" >> $GITHUB_ENV
           CUDA_CORE_ARTIFACT_BASENAME="cuda-core-python${PYTHON_VERSION_FORMATTED}-${{ matrix.host-platform }}"
           echo "CUDA_CORE_ARTIFACT_BASENAME=${CUDA_CORE_ARTIFACT_BASENAME}" >> $GITHUB_ENV
           echo "CUDA_CORE_ARTIFACT_NAME=${CUDA_CORE_ARTIFACT_BASENAME}-${{ github.sha }}" >> $GITHUB_ENV
@@ -139,10 +139,10 @@ jobs:
           # CIBW mounts the host filesystem under /host
           CIBW_ENVIRONMENT_LINUX: >
             CUDA_PATH=/host/${{ env.CUDA_PATH }}
-            PARALLEL_LEVEL=${{ env.PARALLEL_LEVEL }}
+            CUDA_BINDINGS_PARALLEL_LEVEL=${{ env.CUDA_BINDINGS_PARALLEL_LEVEL }}
           CIBW_ENVIRONMENT_WINDOWS: >
             CUDA_HOME="$(cygpath -w ${{ env.CUDA_PATH }})"
-            PARALLEL_LEVEL=${{ env.PARALLEL_LEVEL }}
+            CUDA_BINDINGS_PARALLEL_LEVEL=${{ env.CUDA_BINDINGS_PARALLEL_LEVEL }}
         with:
           package-dir: ./cuda_bindings/
           output-dir: ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }}
@@ -157,10 +157,9 @@ jobs:
           $CHOWN -R $(whoami) ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }}
           ls -lahR ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }}
 
-      # TODO: enable this after NVIDIA/cuda-python#297 is resolved
-      # - name: Check cuda.bindings wheel
-      #   run: |
-      #     twine check ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }}/*.whl
+      - name: Check cuda.bindings wheel
+        run: |
+          twine check ${{ env.CUDA_BINDINGS_ARTIFACTS_DIR }}/*.whl
 
       - name: Upload cuda.bindings build artifacts
         uses: actions/upload-artifact@v4
@@ -221,7 +220,7 @@ jobs:
           - "3.9"
         cuda-version:
           # Note: this is for test-time only.
-          - "12.6.2"
+          - "12.8.0"
           - "12.0.1"
           - "11.8.0"
         local-ctk:
@@ -236,7 +235,7 @@ jobs:
         include:
           - host-platform: linux-64
             python-version: "3.12"
-            cuda-version: "12.6.2"
+            cuda-version: "12.8.0"
             local-ctk: 1
             runner: H100
     name: Test (${{ matrix.host-platform }}, Python ${{ matrix.python-version }}, CUDA ${{ matrix.cuda-version }}, Runner ${{ matrix.runner }}, ${{ (matrix.local-ctk == '1' && 'local CTK') || 'CTK wheels' }})
@@ -263,6 +262,7 @@ jobs:
     permissions:
       id-token: write
       contents: write
+      pull-requests: write
     needs:
       - build
     secrets: inherit