Tags: NVIDIA/cuda-python
Tags
Prepare cuda.core v0.7.0 release (#1877) * Prepare cuda.core v0.7.0 release Finalize release notes with all changes since v0.6.0: - Explicit graph construction (GraphDef, GraphBuilder, typed nodes) - CUDA-Graphics (OpenGL) interop via GraphicsResource - TensorMapDescriptor for Hopper+ TMA - StridedMemoryView DLPack export and C exchange API - NVRTC PCH runtime APIs on Program - CPU callbacks for stream capture (GraphBuilder.callback) - CUDA 13.2 support - Multiple bug fixes and enhancements Also: - Add 0.7.0 to nv-versions.json - Bump pixi.toml version to 0.7.0 - Add GraphicsResource, TensorMapDescriptor to api.rst - Remove "(experimental)" from pyproject.toml and README.md Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Fix install page URL in cuda_core README Point to cuda-core's own install page instead of cuda-bindings. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Address review feedback on release notes - Use consistent "CUDA-OpenGL" naming (not "CUDA-Graphics") - Highlight DLPack export via from_dlpack() array API; move C exchange API detail to New features section - TensorMapDescriptor: reference public StridedMemoryView.as_tensor_map() instead of private _from_tiled/_from_im2col methods Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Update cuda_core/docs/source/release/0.7.0-notes.rst * Trim DLPack export bullet per review suggestion Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Update cuda_core/docs/source/release/0.7.0-notes.rst * Address second round of release note review feedback - Fix PCH entry: reference actual public API (ProgramOptions fields, Program.pch_status property) instead of non-existent methods - Combine ManagedMemoryResource NUMA entries into single bullet - Combine PinnedMemoryResource NUMA entries into single bullet - Replace :issue: role (not configured) with explicit GitHub links - Use :class: cross-ref for ManagedMemoryResource in fixes section Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Use :meth: cross-ref for Program.compile in PCH entry Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Use cyclass.rst template for Program, Linker, ObjectCode, Kernel These Cython classes were using the default autosummary template, which does not expand methods and properties. Switch to cyclass.rst so that properties like Program.pch_status and methods like Program.compile appear in the generated docs and can be cross-referenced. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * remove incorrect entry slipping for --------- Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]>
[Backport] #1606 & #873 (#1750) * ci: require tag-triggered artifacts for release uploads (#1606) Backport of #1606 to 12.9.x. Adapted for the 12.9.x branch workflow structure: - ci.yml: add tag push triggers (v*, cuda-core-v*, cuda-pathfinder-v*) so setuptools-scm resolves exact release versions from tag refs - release.yml: make run-id optional with auto-detection from tag-triggered CI runs via ci/tools/lookup-run-id; add wheel version validation via ci/tools/validate-release-wheels before publishing - Add ci/tools/lookup-run-id: finds the successful tag-triggered CI run for a given release tag - Add ci/tools/validate-release-wheels: rejects dev/local wheel versions and enforces version match against the release tag - release_checklist.yml: add reminder to wait for tag-triggered CI * ci: sync release-upload.yml with main and add download-wheels script Sync release-upload.yml to match main by adding run-id and component inputs and a 'Download and Upload Wheels' step that downloads wheels via ci/tools/download-wheels, validates them via validate-release-wheels, and uploads them to the GitHub Release. Also add the ci/tools/download-wheels helper script (from main) and wire up release.yml to pass run-id and component to the upload-archive job. Fixes #1120: the download-wheels script now exists on the backport branch, so the release-upload workflow no longer fails with 'No such file or directory'. * [pre-commit.ci] auto code formatting --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
[doc-only] Prepare for cuda-pathfinder v1.4.1 release (#1728) * Add Linux support for loading libcupti.so.12 and libcupti.so.13 This commit adds support for finding and loading CUPTI libraries on Linux through cuda.pathfinder. It implements support for all enumerated installation methods: - Site-packages: nvidia/cuda_cupti/lib (CUDA 12) and nvidia/cu13/lib (CUDA 13) - Conda: $CONDA_PREFIX/lib (colocated with other CUDA libraries) - CTK via CUDA_HOME: $CUDA_HOME/extras/CUPTI/lib64 - CTK via canary probe: system CTK root discovery (similar to nvvm) Changes: - Add 'cupti' to supported library names and SONAMEs - Add site-packages paths for CUDA 12 and 13 - Add cupti to CTK root canary discoverable libraries - Update find_nvidia_dynamic_lib to handle extras/CUPTI/lib64 path - Add logic to distinguish CTK (extras/CUPTI/lib64) vs conda (lib) paths - Update _find_so_using_lib_dir to support versioned libraries via glob - Add comprehensive mock tests covering all installation methods Fixes #1572 (Linux support) Made-with: Cursor * Update cupti tests to use new SearchContext-based API Migrated test_load_nvidia_dynamic_lib_using_mocker.py from the old _FindNvidiaDynamicLib API to the new descriptor-based SearchContext API. Changes: - Replace _FindNvidiaDynamicLib imports with search_steps and load_nvidia_dynamic_lib modules - Update mocks to use run_find_steps, LOADER, and SearchContext - Use LIB_DESCRIPTORS to get cupti descriptor - Update all test functions to work with the new search step architecture Made-with: Cursor * Remove unused CTK canary variables from supported_nvidia_libs.py These variables (_CTK_ROOT_CANARY_ANCHOR_LIBNAMES and _CTK_ROOT_CANARY_DISCOVERABLE_LIBNAMES) were added in the cupti PR but are not used in the new descriptor-based architecture. The new code uses desc.ctk_root_canary_anchor_libnames directly from descriptors. Made-with: Cursor * Improve comment for change in LinuxSearchPlatform.find_in_lib_dir() * Add cputi to cu12, cu13 groups in cuda_pathfinder/pyproject.toml * Add cuda_cupti to cuda-components in .github/actions/fetch_ctk/action.yml * Add windows_dlls, site_packages_windows, anchor_rel_dirs_windows for cupti in /descriptor_catalog.py * test: Refactor cupti mock tests to focus on Conda and error paths Remove tests covered by real CI: - Site-packages tests (CUDA 12 and 13) - covered by real CI - CTK tests (CUDA_HOME and canary probe) - covered by real CI - Search order tests involving site-packages/CTK - covered by real CI Keep tests not covered by real CI: - Conda discovery test - Conda not covered by real CI - Error path test (not found) - error path not covered - Conda vs CTK search order test - Conda not covered by real CI Also remove unused imports and helper functions. Made-with: Cursor * Add pathfinder release/1.4.1-notes.rst * Add PR #1731 to release/1.4.1-notes.rst
PreviousNext