Drop `experimental` from cuda.core namespace imports #676

brandon-b-miller · 2025-12-18T18:32:08Z

This PR drops the experimental part of the cuda.core namespace.

copy-pr-bot · 2025-12-18T18:32:11Z

Auto-sync is disabled for ready for review pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

brandon-b-miller · 2025-12-18T18:32:56Z

/ok to test

greptile-apps · 2025-12-18T18:35:19Z

Greptile Summary

This PR introduces a backward-compatible transition from cuda.core.experimental to cuda.core namespace by implementing a compatibility layer.

Key Changes:

Created numba_cuda/numba/cuda/_compat.py that conditionally imports from cuda.core.experimental (for cuda-core <0.5.0) or cuda.core (for >=0.5.0)
Updated all direct imports of cuda.core.experimental across production code and tests to use numba.cuda._compat
Added packaging dependency to support version parsing
Pinned pixi to v0.61.0 in CI workflows

Implementation Quality:

Clean abstraction that centralizes version-specific logic in one file
All imports consistently updated across the codebase
No remaining direct references to cuda.core.experimental outside the compatibility layer

Confidence Score: 5/5

This PR is safe to merge with minimal risk
The implementation is well-structured with a clean compatibility layer that handles both old and new cuda-core versions. All imports are systematically updated, dependencies properly added, and the approach is non-breaking for existing users.
No files require special attention

Important Files Changed

Filename	Overview
numba_cuda/numba/cuda/_compat.py	New compatibility layer added to conditionally import from `cuda.core.experimental` (<0.5.0) or `cuda.core` (>=0.5.0)
numba_cuda/numba/cuda/cudadrv/driver.py	Updated imports to use `numba.cuda._compat` instead of direct `cuda.core.experimental` imports
numba_cuda/numba/cuda/dispatcher.py	Updated imports to use `numba.cuda._compat` for launch and LaunchConfig
pyproject.toml	Added `packaging` dependency to support version parsing in compatibility layer
conda/recipes/numba-cuda/meta.yaml	Added `packaging` to runtime dependencies

greptile-apps

Additional Comments (1)

numba_cuda/numba/cuda/dispatcher.py, line 44 (link)

syntax: Missed import: LaunchConfig still imported from cuda.core.experimental

_{8 files reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

brandon-b-miller · 2025-12-18T19:45:50Z

/ok to test

greptile-apps

Additional Comments (2)

numba_cuda/numba/cuda/_compat.py, line 8-11 (link)

logic: empty if/else blocks cause all imports from _compat to fail - nothing is re-exported. Add imports in each branch:
numba_cuda/numba/cuda/tests/cudadrv/test_cuda_driver.py, line 19 (link)

style: inconsistent with other test files - test_events.py imports Device from _compat while this imports directly from cuda.core

_{Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!}

_{11 files reviewed, 2 comments}

_{Edit Code Review Agent Settings | Greptile}

brandon-b-miller · 2025-12-18T20:14:41Z

/ok to test

greptile-apps

Additional Comments (4)

numba_cuda/numba/cuda/_compat.py, line 31-41 (link)

logic: ObjectCode and LaunchConfig are missing from this compatibility module but are imported from it in several files (driver.py, dispatcher.py, nrt.py, test_module_callbacks.py, test_compiler.py). This will cause ImportError at runtime.
numba_cuda/numba/cuda/_compat.py, line 8-18 (link)

logic: ObjectCode and LaunchConfig need to be imported in both version branches since they're used by downstream modules.
numba_cuda/numba/cuda/_compat.py, line 19-29 (link)

logic: ObjectCode and LaunchConfig need to be imported in both version branches since they're used by downstream modules.
numba_cuda/numba/cuda/tests/cudadrv/test_cuda_driver.py, line 19 (link)

logic: Importing Device directly from cuda.core bypasses the version compatibility layer. This will break with cuda-core < 0.5.0 where Device is only available in cuda.core.experimental.

_{14 files reviewed, 4 comments}

_{Edit Code Review Agent Settings | Greptile}

brandon-b-miller · 2025-12-18T21:03:02Z

/ok to test

brandon-b-miller · 2025-12-18T21:48:34Z

/ok to test ec8d388

bdice

Looks right to me!

bdice · 2025-12-18T22:49:18Z

numba_cuda/numba/cuda/cudadrv/driver.py

    Stream as ExperimentalStream,
    Device as ExperimentalDevice,


You could drop "Experimental" from these names.

I agree with this, however there's some broader work to be done to disambiguate the cuda.core classes with the ones defined later in this file, so I'm going to leave it for a follow up for now.

Backport #676

Stopgap until NVIDIA/numba-cuda#676 is in. Authors: - Jim Crist-Harif (https://github.com/jcrist) Approvers: - Dante Gama Dessavre (https://github.com/dantegd) - Bradley Dice (https://github.com/bdice) URL: #7629

…#7629) Stopgap until NVIDIA/numba-cuda#676 is in. Authors: - Jim Crist-Harif (https://github.com/jcrist) Approvers: - Dante Gama Dessavre (https://github.com/dantegd) - Bradley Dice (https://github.com/bdice) URL: rapidsai#7629

- Add arch specific target support (NVIDIA#549) - chore: disable `locked` flag to bypass prefix-dev/pixi#5256 (NVIDIA#714) - ci: relock pixi (NVIDIA#712) - ci: remove redundant conda build in ci (NVIDIA#711) - chore(deps): bump numba-cuda version and relock pixi (NVIDIA#707) - Dropping bits in the old CI & Propagating recent changes from cuda-python (NVIDIA#683) - Fix `test_wheel_deps_wheels.sh` to actually uninstall `nvvm` and `nvrtc` packages for CUDA 13 (NVIDIA#701) - perf: remove some exception control flow and buffer-exception penalization for arrays (NVIDIA#700) - perf: let CAI fall through instead of calling from_cuda_array_interface (NVIDIA#694) - chore: perf lint (NVIDIA#697) - chore(deps): bump deps in pixi lockfile (NVIDIA#693) - fix: use freethreading-supported `_PySet_NextItemRef` where possible (NVIDIA#682) - Support python `3.14` (NVIDIA#599) - Remove customized address space tracking and address class emission in debug info (NVIDIA#669) - Drop `experimental` from cuda.core namespace imports (NVIDIA#676) - Remove dangling references to NUMBA_CUDA_ENABLE_MINOR_VERSION_COMPATIBILITY (NVIDIA#675) - Use `rapidsai/sccache` in CI (NVIDIA#674) - chore(dev-deps): remove ipython and pyinstrument (NVIDIA#670) - Set up a new VM-based CI infrastructure (NVIDIA#604)

- Add arch specific target support (#549) - chore: disable `locked` flag to bypass prefix-dev/pixi#5256 (#714) - ci: relock pixi (#712) - ci: remove redundant conda build in ci (#711) - chore(deps): bump numba-cuda version and relock pixi (#707) - Dropping bits in the old CI & Propagating recent changes from cuda-python (#683) - Fix `test_wheel_deps_wheels.sh` to actually uninstall `nvvm` and `nvrtc` packages for CUDA 13 (#701) - perf: remove some exception control flow and buffer-exception penalization for arrays (#700) - perf: let CAI fall through instead of calling from_cuda_array_interface (#694) - chore: perf lint (#697) - chore(deps): bump deps in pixi lockfile (#693) - fix: use freethreading-supported `_PySet_NextItemRef` where possible (#682) - Support python `3.14` (#599) - Remove customized address space tracking and address class emission in debug info (#669) - Drop `experimental` from cuda.core namespace imports (#676) - Remove dangling references to NUMBA_CUDA_ENABLE_MINOR_VERSION_COMPATIBILITY (#675) - Use `rapidsai/sccache` in CI (#674) - chore(dev-deps): remove ipython and pyinstrument (#670) - Set up a new VM-based CI infrastructure (#604)

drop experimental namespace

2448499

brandon-b-miller changed the title ~~drop experimental namespace~~ Drop experimental from cuda.core namespace imports Dec 18, 2025

greptile-apps bot reviewed Dec 18, 2025

View reviewed changes

LaunchConfig

45ca9c1

bdice mentioned this pull request Dec 18, 2025

Update cuda.core imports from experimental to stable rapidsai/rmm#2187

Merged

_compat.py

935a757

greptile-apps bot reviewed Dec 18, 2025

View reviewed changes

fixes

f9fd418

greptile-apps bot reviewed Dec 18, 2025

View reviewed changes

brandon-b-miller added 2 commits December 18, 2025 12:19

remove core

da319b0

extra objects

4a7d735

jcrist mentioned this pull request Dec 18, 2025

Don't fail on cuda.core.experimental deprecation warnings rapidsai/cuml#7629

Merged

pin pixi to 0.61.0

ec8d388

bdice approved these changes Dec 18, 2025

View reviewed changes

brandon-b-miller merged commit 5ad2291 into NVIDIA:main Dec 18, 2025
169 of 170 checks passed

brandon-b-miller mentioned this pull request Dec 18, 2025

Backport 676 To 0.22 #678

Merged

brandon-b-miller added a commit that referenced this pull request Dec 19, 2025

Backport 676 To 0.22 (#678)

5d41fa8

Backport #676

gmarkall mentioned this pull request Jan 12, 2026

Bump version to 0.24.0 #716

Merged

isVoid mentioned this pull request Feb 11, 2026

Create Highlevel Bindings for FP8 Datatype #788

Open

Drop experimental from cuda.core namespace imports #676

Drop experimental from cuda.core namespace imports #676

Uh oh!

Conversation

brandon-b-miller commented Dec 18, 2025

Uh oh!

copy-pr-bot bot commented Dec 18, 2025

Uh oh!

brandon-b-miller commented Dec 18, 2025

Uh oh!

greptile-apps bot commented Dec 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Greptile Summary

Confidence Score: 5/5

Important Files Changed

Uh oh!

greptile-apps bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Additional Comments (1)

Uh oh!

brandon-b-miller commented Dec 18, 2025

Uh oh!

greptile-apps bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Additional Comments (2)

Uh oh!

brandon-b-miller commented Dec 18, 2025

Uh oh!

greptile-apps bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Additional Comments (4)

Uh oh!

brandon-b-miller commented Dec 18, 2025

Uh oh!

brandon-b-miller commented Dec 18, 2025

Uh oh!

bdice left a comment

Choose a reason for hiding this comment

Uh oh!

bdice Dec 18, 2025

Choose a reason for hiding this comment

Uh oh!

brandon-b-miller Dec 18, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Drop `experimental` from cuda.core namespace imports #676

Drop `experimental` from cuda.core namespace imports #676

greptile-apps bot commented Dec 18, 2025 •

edited

Loading

greptile-apps bot left a comment •

edited

Loading

greptile-apps bot left a comment •

edited

Loading

greptile-apps bot left a comment •

edited

Loading