NVIDIA / NVIDIA/cutlass

[QST] Wrong comment of LayoutType::B128_BASE32B in make_umma_desc?

Open Beginner friendly
#3,089 2 comments 0 reactions 0 assignees View on GitHub

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

In the comment of make_umma_desc, with Major::MN, the comment says
/* In units of uint128_t, each UmmaDescriptor Major-MN describes a canonical layout of the form
*
* LayoutType::INTERLEAVE : Swizzle<0,4,3> o smem_ptr o ((1,n),(8,k)):((X,SBO),(1,LBO))
* LayoutType::B32 : Swizzle<1,4,3> o smem_ptr o ((2,n),(8,k)):((1,LBO),(2,SBO))
* LayoutType::B64 : Swizzle<2,4,3> o smem_ptr o ((4,n),(8,k)):((1,LBO),(4,SBO))
* LayoutType::B128 : Swizzle<3,4,3> o smem_ptr o ((8,n),(8,k)):((1,LBO),(8,SBO))
* LayoutType::B128_BASE32B : Swizzle<2,5,2> o smem_ptr o ((8,n),(4,k)):((1,LBO),(4,SBO))
*/
But for LayoutType::B128_BASE32B, the stride_10 should be 8, in units of uint128_t?

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start at the make_umma_desc definition and compare the documented LayoutType::B128_BASE32B stride with the surrounding layout descriptions and units. Update the comment if the reported stride is incorrect, then verify that the documentation matches the intended descriptor layout.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
documentation
Issue type
Documentation
Difficulty
1/5
Estimated time
Under an hour
Activity status
Quiet
Clarity
Clearly specified
Newbie friendliness
70/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.