Optimize hip multidimensional kernels
Open
Nobody has claimed this yet.
hip
- Dominant language
- Jupyter Notebook
- Stars
- 135
- Forks
- 55
- Avg merge
- 4d 16h
- Merged PRs (30d)
- 5
Description
The hip multidimensional kernels are currently copies of the cuda variants and use 32 threads in the x block dimension. We should look at using 64 threads instead as that matches the hip wavefront size.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Locate the HIP multidimensional kernels and compare them with their CUDA variants, focusing on the x block dimension and the stated 32-thread configuration. Check how the relevant performance measurements and correctness checks are run. Done means evaluating the 64-thread configuration while preserving kernel correctness and documenting the performance result.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- performance
- Issue type
- Refactor
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100