Skip to content

[merge after release] Restrict PTX 88 to sm_121 for CUDA 12.9+#250

Merged
phiandark merged 1 commit intomainfrom
fix-ptx88-sm121
Mar 30, 2026
Merged

[merge after release] Restrict PTX 88 to sm_121 for CUDA 12.9+#250
phiandark merged 1 commit intomainfrom
fix-ptx88-sm121

Conversation

@hsadasivan
Copy link
Copy Markdown
Collaborator

PTX version 88 breaks other configurations with CUDA 12.9. This change ensures PTX 88 is only used for GB10/sm_121 (DGX Spark), while other architectures use the safer PTX 87.

@copy-pr-bot
Copy link
Copy Markdown

copy-pr-bot bot commented Feb 9, 2026

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.

hsadasivan added a commit that referenced this pull request Feb 9, 2026
…notes

Explains that PTX 87 is used in this release and provides guidance for users
needing DGX Spark/GB10 support with CUDA 12.9 via PR #250.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
@hsadasivan hsadasivan requested a review from phiandark February 9, 2026 22:08
@mariogeiger
Copy link
Copy Markdown
Collaborator

@hsadasivan can you check the conflicts?

PTX version 88 breaks other configurations with CUDA 12.9. This change
ensures PTX 88 is only used for GB10/sm_121 (DGX Spark), while other
architectures use the safer PTX 87.

Co-Authored-By: Hari Sadasivan <hsadasivan@nvidia.com>
@hsadasivan
Copy link
Copy Markdown
Collaborator Author

Done, @mariogeiger

@phiandark phiandark merged commit 90de69f into main Mar 30, 2026
9 checks passed
@mariogeiger mariogeiger deleted the fix-ptx88-sm121 branch March 30, 2026 18:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants