Add support for (gpu-to-gpu) deterministic outputs to multi-cta version of `DeviceBatchedTopK`
Open
- Dominant language
- C++
- Stars
- 2.5k
- Forks
- 486
- Avg merge
- 2d 6h
- Merged PRs (30d)
- 295
Description
This issue has no description.
Contributor guide
Research direction
Start by locating the multi-CTA implementation and tests for DeviceBatchedTopK. Read how GPU-to-GPU output comparisons are currently handled, then determine the expected deterministic behavior and add coverage for it. Done means the multi-CTA version produces deterministic outputs across GPU-to-GPU runs.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- hpc, performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100