Skip to content

WIP: Make sure TMA is compiled in (for now, see if this fails CI) - #2428

Draft
seberg wants to merge 2 commits into
NVIDIA:mainfrom
seberg:tensor-map-linking
Draft

WIP: Make sure TMA is compiled in (for now, see if this fails CI)#2428
seberg wants to merge 2 commits into
NVIDIA:mainfrom
seberg:tensor-map-linking

Conversation

@seberg

@seberg seberg commented Jul 27, 2026

Copy link
Copy Markdown
Contributor

In my conda setup, compiling tensor_map created linking issues and it seems like the reason this didn't fail here might be that it is missingi when it shouldn't be.

(Once/if this fails, let's fix it/)

@copy-pr-bot

copy-pr-bot Bot commented Jul 27, 2026

Copy link
Copy Markdown
Contributor

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.

@github-actions github-actions Bot added the cuda.core Everything related to the cuda.core module label Jul 27, 2026
@seberg

seberg commented Jul 27, 2026

Copy link
Copy Markdown
Contributor Author

/ok to test 2c99b8a

seberg added 2 commits July 27, 2026 15:41
In my conda setup, compiling `tensor_map` created linking issues and
it seems like the reason this didn't fail here might be that it is
missing when it shouldn't be.

Adding the `cccl` headers breaks CI the same way, but moving things
to `_tensor_map_cccl.pxd` and soft-linking it should then unbreak
things again.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

cuda.core Everything related to the cuda.core module

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant