cuda.core v1.0.0 release prep: support policy, release notes, doc cleanup by leofang · Pull Request #2032 · NVIDIA/cuda-python

leofang · 2026-05-06T02:38:44Z

Summary

Umbrella PR for cuda.core v1.0.0 release documentation.

Support policy (cuda_core/docs/source/support.rst, new)

SemVer versioning scheme
CUDA version support (dual major versions — currently 12 and 13)
Python version support (CPython EOL schedule)
Project lifecycle and release cadence (bimonthly target, no backport policy, deprecation notice policy)
Experimental free-threading build support
Note on minimum cuda-bindings / driver version requirements for certain features

Release notes (cuda_core/docs/source/release/1.0.0-notes.rst)

Filled in all missing entries since cuda-core-v0.7.0: new features (green contexts, program cache, system NVML APIs, NVML enum re-wrapping, StrEnum support), breaking changes (tensor bridge behavior, experimental namespace removal, graph API surface trim, system.Device renames, driver version API consolidation, explicit stream requirement, Linker.which_backend, peer_accessible_by MutableSet, and more), and bug fixes (is_managed, Device.arch, nvJitLink logs, DeviceEvents/SystemEvents dealloc, graph kernel-argument lifetime, runtime error messages, field values, wheel size)
Added Highlights section

Doc updates

api.rst: replaced pre-1.0 "APIs may change" warning with stable-API statement linking to support policy
api_nvml.rst (new): split cuda.core.system API reference into its own page
install.rst: updated free-threading version reference from 0.4.0 to 1.0.0
nv-versions.json: added 1.0.0 entry for version switcher dropdown
index.rst: added support and api_nvml to toctree

Removed deprecated cuda.core.experimental namespace

Deleted cuda_core/cuda/core/experimental/ shim and test_experimental_backward_compat.py

README / PyPI description fixes (across README.md, cuda_python/DESCRIPTION.rst, cuda_python/docs/source/index.rst)

Fixed broken CCCL doc URLs for cuda.coop and cuda.compute (old paths return 404)
Added missing nvFatbin and NVML to the cuda.bindings interface list

Misc

Removed hash algorithm and digest size from make_program_cache_key public docstring (implementation detail, see make_program_cache_key: blake2b is not a FIPS-compliant hash #2043)

Closes#2030

Add support.rst covering versioning (SemVer), CUDA version support (dual major versions), Python version support (CPython EOL schedule), free-threading (experimental), and release cadence (bimonthly). ClosesNVIDIA#2030

copy-pr-bot · 2026-05-06T02:38:48Z

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

- Update cuda.coop and cuda.compute URLs from the old nvidia.github.io/cccl/python/{coop,compute} paths (now 404) to the current unstable doc paths. - Add nvFatbin and NVML to the cuda.bindings interface list. - Update all three synced files: README.md, cuda_python/DESCRIPTION.rst, and cuda_python/docs/source/index.rst.

Add new features (green contexts, system.Device NVML APIs, system.typing module, NVML enum re-wrapping), breaking changes (tensor bridge behavior, system.Device renames, privatized helper classes, UUID format change, removed enums), and bug fixes (is_managed for pool alloc, nvJitLink log error handling, NVML event set init, Device.arch unknown, empty field values, runtime error messages, wheel size reduction).

- api.rst: replace pre-1.0 warning with stable-API statement and link to support policy. - install.rst: update free-threading version reference from 0.4.0 to 1.0.0. - nv-versions.json: add 1.0.0 entry for the version switcher dropdown.

# Conflicts: # cuda_core/docs/source/release/1.0.0-notes.rst

Move the CUDA system information / NVML section from api.rst into a dedicated api_nvml.rst. The new page uses its own `.. module:: cuda.core.system` directive so autosummary entries no longer need the `system.` prefix. Added to index.rst toctree after api.

leofang · 2026-05-06T17:48:05Z

FYI @rparolin @danielfrg @kkraus14 plz review the support policy, which is written based on the signed-off GA version scheme

danielfrg · 2026-05-06T19:20:17Z

lgtm

The Returns section exposed the hash algorithm and digest size, which are implementation details. Replace with "opaque bytes digest" so the public API contract does not pin these. See NVIDIA#2043

# Conflicts: # cuda_core/docs/source/api.rst

The cuda.core.experimental namespace was deprecated in v0.5.0 when all public APIs moved to the top-level cuda.core namespace. Remove the backward-compatibility shim and its test as promised for v1.0.0.

leofang · 2026-05-07T17:28:22Z

/ok to test da45345

# Conflicts: # cuda_core/cuda/core/experimental/__init__.py # cuda_core/docs/source/release/1.0.0-notes.rst # cuda_core/tests/test_experimental_backward_compat.py

…#2047 - New feature: persistent program cache for Program.compile (InMemoryProgramCache, FileStreamProgramCache, make_program_cache_key). - Fix: graph kernel nodes now prevent kernel-argument GC. - Fix: DeviceEvents.__dealloc__ crash on uninitialized handle.

leofang · 2026-05-07T22:10:32Z

/ok to test 6eb0a45

rparolin · 2026-05-07T22:23:07Z

Code review

Found 1 issue:

cuda_core/docs/source/install.rst rewords "As of cuda-core 0.4.0" to "As of cuda-core 1.0.0" for the free-threaded build note. Free-threaded wheels were introduced in 0.4.0 (see docs: add FT build notes #1071, commit 3bf736b), not 1.0.0 — this regresses the historical accuracy of when FT support became available. The original "0.4.0" should be retained.

cuda-python/cuda_core/docs/source/install.rst

Lines 32 to 37 in 6eb0a45

 Free-threading Build Support 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 As of cuda-core 1.0.0, **experimental** packages for the `free-threaded interpreter`_ are shipped. 
 1. Support for these builds is best effort, due to heavy use of `built-in

🤖 Generated with Claude Code

_{- If this code review was useful, please react with 👍. Otherwise, react with 👎.}

leofang · 2026-05-07T23:30:24Z

Admin-merging since Mike's requests already addressed yesterday and we have a tight schedule coupled to CI instability. I need to kick off the release process asap.

github-actions · 2026-05-08T00:31:26Z

Doc Preview CI
Preview removed because the pull request was closed or merged.

Document cuda.core support policy
1eeb937
Add support.rst covering versioning (SemVer), CUDA version support (dual major versions), Python version support (CPython EOL schedule), free-threading (experimental), and release cadence (bimonthly). ClosesNVIDIA#2030

github-actionsBot added the cuda.core Everything related to the cuda.core module label May 6, 2026

leofang added this to the cuda.core v1.0.0 milestone May 6, 2026

leofang self-assigned this May 6, 2026

leofang added documentation Improvements or additions to documentation P1 Medium priority - Should do labels May 6, 2026

leofang changed the title ~~Document cuda.core support policy~~Document cuda.core support policy and fix READMEMay 6, 2026

leofang added 2 commits May 6, 2026 03:31

Update cuda.core docs for 1.0.0 GA
1878060
- api.rst: replace pre-1.0 warning with stable-API statement and link to support policy. - install.rst: update free-threading version reference from 0.4.0 to 1.0.0. - nv-versions.json: add 1.0.0 entry for the version switcher dropdown.

leofang changed the title ~~Document cuda.core support policy and fix README~~cuda.core 1.0.0 release prep: support policy, docs, and release notesMay 6, 2026

leofang added P0 High priority - Must do! and removed P1 Medium priority - Should do labels May 6, 2026

mdboom requested changes May 6, 2026
View reviewed changes

mdboom reviewed May 6, 2026
View reviewed changes

Comment threadcuda_core/docs/source/release/1.0.0-notes.rst Outdated

leofang added 2 commits May 6, 2026 17:28

Merge remote-tracking branch 'origin/main' into cuda-core-support-policy
9b1efcf
# Conflicts: # cuda_core/docs/source/release/1.0.0-notes.rst

leofang mentioned this pull request May 6, 2026
make_program_cache_key: blake2b is not a FIPS-compliant hash #2043
Closed

leofang added 2 commits May 6, 2026 22:55

Merge remote-tracking branch 'origin/main' into cuda-core-support-policy
c6e413b

Remove algorithm and size details from make_program_cache_key docstring
b08abc4
The Returns section exposed the hash algorithm and digest size, which are implementation details. Replace with "opaque bytes digest" so the public API contract does not pin these. See NVIDIA#2043

kkraus14 reviewed May 6, 2026
View reviewed changes

Comment threadcuda_core/docs/source/support.rst Outdated

leofang added 2 commits May 7, 2026 16:14

Merge remote-tracking branch 'origin/main' into cuda-core-support-policy
096eb46
# Conflicts: # cuda_core/docs/source/api.rst

Remove deprecated cuda.core.experimental namespace
da45345
The cuda.core.experimental namespace was deprecated in v0.5.0 when all public APIs moved to the top-level cuda.core namespace. Remove the backward-compatibility shim and its test as promised for v1.0.0.

leofang commented May 7, 2026
View reviewed changes

Comment threadcuda_core/docs/source/support.rst

leofang commented May 7, 2026
View reviewed changes

Comment threadcuda_core/docs/source/support.rst

Merge remote-tracking branch 'origin/main' into cuda-core-support-policy
8b4852b
# Conflicts: # cuda_core/cuda/core/experimental/__init__.py # cuda_core/docs/source/release/1.0.0-notes.rst # cuda_core/tests/test_experimental_backward_compat.py

leofang added 3 commits May 7, 2026 21:38

Update 1.0.0-notes.rst
9a1e4ba

expand support policy
6eb0a45

leofang marked this pull request as ready for review May 7, 2026 22:10

leofang changed the title ~~cuda.core 1.0.0 release prep: support policy, docs, and release notes~~cuda.core v1.0.0 release prep: support policy, release notes, doc cleanupMay 7, 2026

leofang requested review from Andy-Jost, danielfrg, kkraus14, mdboom and rparolin May 7, 2026 22:15

This comment has been minimized.
Sign in to view

rparolin reviewed May 7, 2026
View reviewed changes

Comment threadcuda_core/docs/source/install.rst Outdated

danielfrg approved these changes May 7, 2026
View reviewed changes

rparolin approved these changes May 7, 2026
View reviewed changes

leofang commented May 7, 2026
View reviewed changes

Comment threadcuda_core/docs/source/install.rst Outdated
Comment threadcuda_core/docs/source/support.rst Outdated
Comment threadcuda_core/docs/source/install.rst Outdated

wordsmith
80d27a3

leofang merged commit 9cc3420 into NVIDIA:mainMay 7, 2026
4 checks passed

leofang deleted the cuda-core-support-policy branch May 7, 2026 23:30

This comment has been minimized.
Sign in to view

aryanputta mentioned this pull request May 14, 2026
Use FIPS-safe hashes for program cache keys #2087
Merged

Uh oh!

Conversation

leofang commented May 6, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Uh oh!

copy-pr-botBot commented May 6, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

leofang commented May 6, 2026

Uh oh!

danielfrg commented May 6, 2026

Uh oh!

Uh oh!

leofang commented May 7, 2026

Uh oh!

Uh oh!

Uh oh!

leofang commented May 7, 2026

Uh oh!

rparolin commented May 7, 2026

Code review

Uh oh!

This comment has been minimized.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

leofang commented May 7, 2026

Uh oh!

Uh oh!

This comment has been minimized.

github-actionsBot commented May 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

leofang commented May 6, 2026•
edited
Loading