Skip to content

Expand wheel build to arm64 and Windows - #2387

Draft
clin1234 wants to merge 11 commits into
NVIDIA:mainfrom
clin1234:315-support
Draft

clin1234 wants to merge 11 commits into
NVIDIA:mainfrom
clin1234:315-support

Conversation

@clin1234

@clin1234 clin1234 commented Jul 18, 2026 •

Copy link
Copy Markdown
Contributor

Description

closes #2163

Checklist

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

@copy-pr-bot

copy-pr-bot Bot commented Jul 18, 2026

Copy link
Copy Markdown
Contributor

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.

@github-actions github-actions Bot added CI/CD CI/CD infrastructure cuda.bindings Everything related to the cuda.bindings module cuda.core Everything related to the cuda.core module cuda.pathfinder Everything related to the cuda.pathfinder module labels Jul 18, 2026
@clin1234 clin1234 changed the title 315 support Expand wheel build to arm64 and Windows Jul 18, 2026

@leofang leofang left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for your contribution! I don't think we want to burn resources this much for pre-release software on a per-PR basis. Could you keep 3.15/3.15t to nightly-standard rows only?

@leofang leofang added P2 Low priority - Nice to have and removed cuda.bindings Everything related to the cuda.bindings module cuda.core Everything related to the cuda.core module cuda.pathfinder Everything related to the cuda.pathfinder module labels Jul 20, 2026
@github-actions github-actions Bot added cuda.bindings Everything related to the cuda.bindings module cuda.core Everything related to the cuda.core module cuda.pathfinder Everything related to the cuda.pathfinder module labels Jul 20, 2026
Removed configurations for Python version 3.15 and 3.15t from the test matrix.
Comment thread cuda_pathfinder/pyproject.toml Outdated
host = [
# TODO: remove the Python 3.15 guard once 3.15 is officially supported
"nvidia-cutlass; python_version < '3.15'",
"nvidia-cutlass",

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Did you check that cutlass already has 3.15 support?

Comment thread cuda_core/pyproject.toml
Comment thread cuda_core/pyproject.toml Outdated
Comment thread cuda_bindings/pyproject.toml
Comment thread ci/tools/download-wheels
Comment thread ci/test-matrix.yml Outdated
Comment thread ci/test-matrix.yml Outdated
Comment thread ci/test-matrix.yml Outdated
Comment thread ci/test-matrix.yml Outdated
clin1234 and others added 4 commits July 20, 2026 14:32
Co-authored-by: Leo Fang <leo80042@gmail.com>
Added new configurations for Python 3.15 and 3.15t with CUDA 13.3.0 for AMD64 and ARM64 architectures.
@clin1234

Copy link
Copy Markdown
Contributor Author

@leofang, please review

@clin1234
clin1234 requested a review from leofang October 11, 2026 09:21
@clin1234
clin1234 marked this pull request as ready for review October 11, 2026 09:23
@clin1234
clin1234 marked this pull request as draft October 11, 2026 09:25
@coderabbitai

coderabbitai Bot commented Oct 11, 2026 •

Copy link
Copy Markdown

Review in Change Stack →

📝 Summary

Summary by CodeRabbit

  • Tests
    • Expanded nightly validation coverage for CUDA 13.4.2 across additional GPU architectures and Python versions, including free-threaded Python builds. The new test configurations cover A100 and L4 GPUs, helping verify compatibility across a wider range of supported environments.

Walkthrough

The Linux nightly-standard CI matrix adds four test entries for Python 3.15 and 3.15t. The SPDX copyright year changes from 2025 to 2026.

Changes

Python 3.15 CI

Layer / File(s) Summary
Add Python 3.15 nightly matrix entries
ci/test-matrix.yml
Adds four Linux nightly-standard entries with CUDA 13.4.2. The arm64 entries use an A100 or L4; the amd64 entries use an A100 and DRIVER_MODE: MCDM. Each entry specifies one GPU and the latest driver.

Assessment against linked issues

Objective Addressed Explanation
Build and test Python 3.15 pre-release in CI, and store artifacts [#2163] ❓ The matrix adds Python 3.15 test entries. The supplied changes do not show whether CI stores artifacts.
Release Python 3.15 wheels and conda packages when Python 3.15.0 is out [#2163] ❌ No wheel or conda package release changes are shown.
Drop Python 3.10 [#2163] ❌ No Python 3.10 removal is shown.

Priority: ➖ Normal

Change: Feature

Merge Risk: 🔵 Low · up to 249d7

The new amd64 Python 3.15 configurations are not exercised by nightly-standard, leaving those combinations without the intended CI coverage. The arm64 entries are covered; add an amd64 job or remove the unused rows.

✨ Finishing Touches 💡 1
🛠️ Fix failing CI checks 💡
  • Commit to this branch
  • Create a new PR

  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Comment @coderabbitai help to get the list of available commands.

@coderabbitai coderabbitai Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actionable comments posted: 1


ℹ️ Review info
⚙️ Run configuration
  • Configuration used: Repository: NVIDIA/cuda-python/.coderabbit.yaml
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: c5d5f34c-1a63-404f-b6a6-5133a9df1594
📥 Commits

Reviewing files that changed from the base of the PR and between c7e6f09 and 249d73c.

📒 Files selected for processing (2)
  • ci/test-matrix.yml
  • ci/tools/download-wheels

Included review availability: This review used your included allowance. Your plan provides up to 8 included reviews per hour; 7 remain after this review.

Comment thread ci/test-matrix.yml
Comment on lines +121 to +122
- { ARCH: 'amd64', PY_VER: '3.15', CUDA_VER: '13.4.2', LOCAL_CTK: '0', GPU: 'a100', GPU_COUNT: '1', DRIVER: 'latest', DRIVER_MODE: 'MCDM', ENV: { MODE: 'nightly-standard' } }
- { ARCH: 'amd64', PY_VER: '3.15t', CUDA_VER: '13.4.2', LOCAL_CTK: '0', GPU: 'a100', GPU_COUNT: '1', DRIVER: 'latest', DRIVER_MODE: 'MCDM', ENV: { MODE: 'nightly-standard' } }

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win

suggestion: Route the amd64 entries to a nightly-standard job.

ci-nightly.yml runs nightly-standard only for linux-aarch64. The shared Linux workflow filters matrix entries by the requested architecture. These amd64 rows are therefore never selected and add no amd64 Python 3.15 or 3.15t test coverage. Add a linux-64 nightly-standard job or remove the unused rows. (raw.githubusercontent.com)

As per path instructions: “focus on matrix correctness” and “targeted build and test behavior.”

Source: Path instructions

@clin1234 clin1234 left a comment

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good

Commented in CodeRabbit Change Stack

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CI/CD CI/CD infrastructure cuda.bindings Everything related to the cuda.bindings module cuda.core Everything related to the cuda.core module cuda.pathfinder Everything related to the cuda.pathfinder module P2 Low priority - Nice to have

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Support Python 3.15

2 participants