NVIDIA / NVIDIA/cutlass

[QST]question about the `sgemm_1.cu`

Open
#1,659 2 comments 1 reaction 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

I'm learning the tutorial about sgemm_1.cu here: https://github.com/NVIDIA/cutlass/blob/v3.5.0/examples/cute/tutorial/sgemm_1.cu

My question is that:
How can we know the output of C?

I see in both cases, the gemm_nt and gemm_tn, the C matrix is always with stide (1, M).

Is that the requirement of cute? Or we can also define it as (M, 1)?

I'm quite confused about this.

Another question, the N, T is based on the condition that a tensor is always column major with out transpose? right?

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 with examples/cute/tutorial/sgemm_1.cu, focusing on the gemm_nt and gemm_tn examples and their C tensor strides. Trace how the tutorial defines the tensor layouts and documents N/T, then clarify how the output layout is determined and whether an alternative layout is supported. Done means the tutorial or its surrounding documentation answers both questions directly.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
documentation
Issue type
Documentation
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.