[QST] Is it possible to have smaller tile shapes such as 128 x 16 for nvfp4 gemms?
Open
Nobody has claimed this yet.
? - Needs Triage
inactive-30d
inactive-90d
question
- Dominant language
- C++
- Stars
- 10.5k
- Forks
- 2.1k
- Avg merge
- 3d 11h
- Merged PRs (30d)
- 7
Description
What is your question?
I'm building a software especially opting for low latency that contains several fp4 gemms.
Unlike normal gemms, fp4 gemms with microscaling types appear to have some limits on their tile shape. How can I get over this limitation?
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
The issue names no files, tests, or entry points. Start by investigating CUTLASS's NVFP4 GEMM tile-shape constraints and how microscaling types impose them. Determine whether a 128 x 16 tile is supported or what limitation prevents it, and make the supported behavior or required change clear.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100