NVIDIA / NVIDIA/cccl

Expose more mbarrier functions in cuda::ptx

Open
#6,236 2 comments 0 reactions 0 assignees View on GitHub
Dominant language
C++
Stars
2.5k
Forks
486
Avg merge
2d 6h
Merged PRs (30d)
295

Description

Reading through `cuda::barrier`, most inline PTX can be replaced by functions from `cuda::ptx` except for these two:

- [x] `mbarrier.inval.shared.b64 [%0];`
- [ ] `mbarrier.arrive_drop.shared.b64 _, [%0];`

We should expose those two as well in `cuda::ptx` so we can remove the inline PTX in `cuda::barrier`.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.