Skip to content

Adopt CUDA stream compatibility accessors - #2557

Open
bdice wants to merge 2 commits into
NVIDIA:mainfrom
bdice:cuda-stream-ref-prep
Open

Adopt CUDA stream compatibility accessors#2557
bdice wants to merge 2 commits into
NVIDIA:mainfrom
bdice:cuda-stream-ref-prep

Conversation

@bdice

@bdice bdice commented Sep 5, 2026

Copy link
Copy Markdown
Contributor

Summary

Use the get() and sync() compatibility aliases added in RMM #2537. These spellings are shared by rmm::cuda_stream_view and cuda::stream_ref.

This preserves existing stream types and public APIs while extracting mechanical accessor updates from the broader stream migration. It is independently buildable without RMM #2372 and leaves the migration PR focused on actual type and signature changes.

This updates raw CUDA, library, kernel-launch, and legacy API boundaries throughout cuVS while preserving current stream types. Changes that require RAFT to return cuda::stream_ref remain in cuVS #2521.

@copy-pr-bot

copy-pr-bot Bot commented Sep 5, 2026

Copy link
Copy Markdown

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.

@bdice

bdice commented Sep 5, 2026

Copy link
Copy Markdown
Contributor Author

Validated this preparation branch against RMM upstream/main: the local -j0 build completed and the focused IVF-PQ test passed. The dependent migration PR #2521 has been rebased onto this branch.

@bdice
bdice marked this pull request as ready for review September 5, 2026 20:02
@bdice
bdice requested review from a team as code owners September 5, 2026 20:02
@bdice bdice added improvement Improves an existing functionality non-breaking Introduces a non-breaking change labels Sep 5, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

improvement Improves an existing functionality non-breaking Introduces a non-breaking change

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant