Skip to content

Conversation

@pentschev
Copy link
Member

No description provided.

@pentschev pentschev self-assigned this Aug 29, 2025
@pentschev pentschev added feature request New feature or request non-breaking Introduces a non-breaking change labels Aug 29, 2025
@copy-pr-bot
Copy link

copy-pr-bot bot commented Aug 29, 2025

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.

@pentschev
Copy link
Member Author

/ok to test

@pentschev pentschev marked this pull request as ready for review September 9, 2025 09:30
@pentschev pentschev requested review from a team as code owners September 9, 2025 09:30
@pentschev pentschev requested a review from a team as a code owner September 9, 2025 11:17
Copy link
Member

@madsbk madsbk left a comment

Choose a reason for hiding this comment

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

Overall looks good

Copy link
Member

@madsbk madsbk left a comment

Choose a reason for hiding this comment

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

Looks good @pentschev, I only have some minor stuff

Copy link
Member Author

@pentschev pentschev left a comment

Choose a reason for hiding this comment

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

Improve formatting

- ${{ compiler("c") }}
- ${{ compiler("cxx") }}
- ${{ compiler("cuda") }}
- cuda-cupti-dev
Copy link
Contributor

Choose a reason for hiding this comment

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

We typically have only compilers in build and all CUDA libraries (e.g. libcublas-dev, cuda-cupti-dev, etc.) in host. (Because the compiler depends on cuda-cudart-dev, it ends up in both build and host.) One way to think about it is that if you were cross-compiling, you'd need your toolchain like the compiler executables in build but all the libraries you're building against go in host. For this reason I suspect we should only need cuda-cupti-dev in host, and that if it isn't working that way, we have some kind of build system bug that needs to be identified and addressed.

For now, I don't want to block further work, but let's flag this:

Suggested change
- cuda-cupti-dev
- cuda-cupti-dev # TODO: This should only be needed in host, and may indicate a packaging issue

Copy link
Member Author

Choose a reason for hiding this comment

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

Following a discussion offline, instead of adding comments we started #510 to track this, given this PR is only an extension of changes already added in #445 , where cuda-cupti-dev is part of build section of librapidsmpf's conda recipe.

@pentschev
Copy link
Member Author

Thanks all for the reviews!

@pentschev
Copy link
Member Author

/merge

@rapids-bot rapids-bot bot merged commit 9b12c6e into rapidsai:branch-25.10 Sep 15, 2025
91 checks passed
@pentschev pentschev deleted the cupti-python branch September 23, 2025 14:11
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

feature request New feature or request non-breaking Introduces a non-breaking change

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants