Use CTK 12.9.1 for `cuda-bindings` 12.9.x by kkraus14 · Pull Request #955 · NVIDIA/cuda-python

kkraus14 · 2025-09-09T21:04:51Z

Description

closes#820

TODO:

Rerun codegen with 12.9.1
Update driver / runtime result explanations
Add release notes

Checklist

New or existing tests cover these changes.
The documentation is up to date with these changes.

copy-pr-bot · 2025-09-09T21:04:54Z

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.

leofang · 2025-09-09T21:07:30Z

ah.. this reminds me, we haven't backported #889 to 12.9.x, should we do it before this PR?

kkraus14 · 2025-09-09T21:20:27Z

ah.. this reminds me, we haven't backported #889 to 12.9.x, should we do it before this PR?

100%. I'll do that as part of this PR.

leofang · 2025-09-09T21:54:00Z

I wonder if the backport bot can handle it

leofang · 2025-09-10T01:49:52Z

I wonder if the backport bot can handle it

Looks like not... #889 (comment)

Let me know if I (or Copilot) should handle this for you.

kkraus14 · 2025-09-10T01:52:25Z

I wonder if the backport bot can handle it
Looks like not... #889 (comment)
Let me know if I (or Copilot) should handle this for you.

I'll take care of it, thanks!

…ith hard-coded values, optimized checkout, and prepared Windows self-hosted runner migration (NVIDIA#889) * Initial plan * Consolidate test matrices from workflows into ci/test-matrix.json Co-authored-by: leofang <5534781+leofang@users.noreply.github.com> * Hard-code all GPU and ARCH values in test-matrix.json with 6 fields per entry Co-authored-by: leofang <5534781+leofang@users.noreply.github.com> * Update Windows test matrix with a100 GPU and latest-1 driver, configure self-hosted runners Co-authored-by: leofang <5534781+leofang@users.noreply.github.com> * fix * Revert eed0b71 and change Windows DRIVER from latest-1 to latest Co-authored-by: leofang <5534781+leofang@users.noreply.github.com> * Add proxy cache setup to Windows workflow for self-hosted runners Co-authored-by: leofang <5534781+leofang@users.noreply.github.com> * Remove Git for Windows and gh CLI installation steps, add T4 GPU support to Windows matrix Co-authored-by: leofang <5534781+leofang@users.noreply.github.com> * Set fetch-depth: 1 for checkout steps and favor L4/T4 over A100 GPUs for Windows testing Co-authored-by: leofang <5534781+leofang@users.noreply.github.com> * Revert Windows workflow to GitHub-hosted runners with TODO comments for future self-hosted migration Co-authored-by: leofang <5534781+leofang@users.noreply.github.com> * [pre-commit.ci] auto code formatting * Revert Win runner name change for now --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: leofang <5534781+leofang@users.noreply.github.com> Co-authored-by: Leo Fang <leof@nvidia.com> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

copy-pr-bot · 2025-09-10T02:08:03Z

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

Contributors can view more details about this message here.

kkraus14 · 2025-09-10T02:11:16Z

/ok to test

kkraus14 · 2025-09-10T03:28:48Z

/ok to test

copy-pr-bot · 2025-09-10T03:28:55Z

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.

cryos · 2025-09-10T16:08:06Z

/ok to test

cryos · 2025-09-10T16:12:06Z

/ok to test

kkraus14 · 2025-09-10T16:14:24Z

 build-type: pull-request
 host-platform: ${{ matrix.host-platform }}
 build-ctk-ver: ${{ needs.ci-vars.outputs.CUDA_BUILD_VER }}
+ matrix_filter: "select(([.CUDA_VER // empty | split(\".\")[] | tonumber] as $v | ($v[0] < 13)))"


nitpick: can we programmatically derive 13 here and so we can have this unified across main and 12.9.x and allow it to nicely carry over to the future where we have main and 13.x.y?

You want to derive the CUDA major version from the branch name? I see the motivation, just thinking through what we want to key from programatically.

Or from the version json file since it would always be aligned to the major version we want to use? It's not a big deal though where if we want to punt to a future PR that's totally fine.

Third time was the charm 🎉 I can look at it for a follow up, it shouldn't be much, but it would be good to get this merged now. Lots of meetings today sadly...

cryos · 2025-09-10T16:56:54Z

/ok to test

cryos · 2025-09-10T19:41:32Z

/ok to test

cryos

CI changes look good to me

kkraus14 · 2025-09-10T20:13:46Z

@cryos I think we need to backport using the self hosted Windows runners as well. Want to do that here or a separate PR?

cryos · 2025-09-10T20:14:58Z

@cryos I think we need to backport using the self hosted Windows runners as well. Want to do that here or a separate PR?

I was about to ask - I think we should, but I was going to do it in a follow up. Happy to do it here if preferred (it is a pretty small patch).

kkraus14 · 2025-09-10T20:18:09Z

@cryos I think we need to backport using the self hosted Windows runners as well. Want to do that here or a separate PR?
I was about to ask - I think we should, but I was going to do it in a follow up. Happy to do it here if preferred (it is a pretty small patch).

If you don't mind lets do it here. Thanks!

Migrate the Windows testing to use the new NV GHA runners. Cherry-pick NVIDIA#958.

cryos · 2025-09-10T20:23:40Z

/ok to test

kkraus14 · 2025-09-10T20:30:09Z

LGTM, thanks @cryos. There's one open question related to cufile that I'd like to hear from @rwgk or @leofang before we merge

cryos · 2025-09-10T20:46:16Z

It looks like everything is passing in CI @kkraus14, it looks good to go from my perspective pending the open questions you mentioned.

leofang

Not sure what the matrix_filter is about (I assume it's to avoid editing the json file?), but LGTM

leofang · 2025-09-11T00:00:49Z

Thanks, Keith, Marcus, Ralf!

bump all CI jobs to CUDA 12.9.1
6d031d9

kkraus14 changed the base branch from main to 12.9.xSeptember 9, 2025 21:05

leofang assigned kkraus14 Sep 10, 2025

leofang added enhancement Any code-related improvements P1 Medium priority - Should do cuda.bindings Everything related to the cuda.bindings module labels Sep 10, 2025

leofang added this to the cuda-python 13-next, 12-next milestone Sep 10, 2025

kkraus14 commented Sep 10, 2025
View reviewed changes

Comment thread.github/workflows/test-wheel-windows.yml Outdated

kkraus14 commented Sep 10, 2025
View reviewed changes

Comment thread.github/workflows/test-wheel-windows.yml Outdated

forgot to add windows
8edaa4d

rerun codegen with 12.9.1 and update result/error explanations
2b6d3ea

kkraus14 commented Sep 10, 2025
View reviewed changes

Comment threadcuda_bindings/cuda/bindings/_internal/cufile.pxd

kkraus14 marked this pull request as ready for review September 10, 2025 03:28

kkraus14 mentioned this pull request Sep 10, 2025
Update ci and bindings to be generated based on CTK 13.0.1 #960
Merged
2 tasks

leofang mentioned this pull request Sep 10, 2025
CI: Move to self-hosted Windows GPU runners #958
Merged
3 tasks

First stab at the filter for CUDA < 13 in CI
c3cef70

cryos force-pushed the ctk_12.9.1 branch from 95a508b to c3cef70Compare September 10, 2025 16:11

kkraus14 commented Sep 10, 2025
View reviewed changes

Get data from the top-level array
1510ff8

Use the map function on select output
18498da

cryos previously approved these changes Sep 10, 2025
View reviewed changes

kkraus14 commented Sep 10, 2025
View reviewed changes

CI: Move to self-hosted Windows GPU runners
90b37d5
Migrate the Windows testing to use the new NV GHA runners. Cherry-pick NVIDIA#958.

cryos dismissed their stale review via 90b37d5September 10, 2025 20:23

leofang approved these changes Sep 10, 2025
View reviewed changes

leofang merged commit 3508f75 into NVIDIA:12.9.xSep 11, 2025
35 checks passed

leofang mentioned this pull request Sep 11, 2025
MNT: Use CTK 12.9.1 to build cuda-bindings 12.9.x #820
Closed

cpcloud mentioned this pull request Oct 8, 2025
backport ft fix #1111
Closed

Uh oh!

Conversation

kkraus14 commented Sep 9, 2025• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Checklist

Uh oh!

copy-pr-botBot commented Sep 9, 2025

Uh oh!

leofang commented Sep 9, 2025

Uh oh!

kkraus14 commented Sep 9, 2025

Uh oh!

leofang commented Sep 9, 2025

Uh oh!

leofang commented Sep 10, 2025

Uh oh!

kkraus14 commented Sep 10, 2025

Uh oh!

copy-pr-botBot commented Sep 10, 2025

Uh oh!

Uh oh!

Uh oh!

kkraus14 commented Sep 10, 2025

Uh oh!

Uh oh!

kkraus14 commented Sep 10, 2025

Uh oh!

copy-pr-botBot commented Sep 10, 2025

Uh oh!

cryos commented Sep 10, 2025

Uh oh!

cryos commented Sep 10, 2025

Uh oh!

kkraus14Sep 10, 2025

Choose a reason for hiding this comment

Uh oh!

cryosSep 10, 2025

Choose a reason for hiding this comment

Uh oh!

kkraus14Sep 10, 2025

Choose a reason for hiding this comment

Uh oh!

cryosSep 10, 2025• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cryos commented Sep 10, 2025

Uh oh!

cryos commented Sep 10, 2025

Uh oh!

cryos left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

kkraus14 commented Sep 10, 2025

Uh oh!

cryos commented Sep 10, 2025

Uh oh!

kkraus14 commented Sep 10, 2025

Uh oh!

cryos commented Sep 10, 2025

Uh oh!

kkraus14 commented Sep 10, 2025

Uh oh!

cryos commented Sep 10, 2025

Uh oh!

leofang left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

leofang commented Sep 11, 2025

Uh oh!

Reviewers

Assignees

Labels

kkraus14 commented Sep 9, 2025•
edited
Loading

cryosSep 10, 2025•
edited
Loading