NVIDIA / NVIDIA/TransformerEngine

[bug] attention benchmark script not work

Open
#1,919 1 comment 0 reactions 1 assignee View on GitHub

@KshitijLakhani is already working on this.

Since Dec 4, 2025.

attention
Dominant language
Python
Stars
3.5k
Forks
831
Avg merge
3d 11h
Merged PRs (30d)
65

Description

I seem to be having some trouble getting the benchmark script to run. I noticed that there have been some API changes, and it looks like the benchmark script hasn't been updated to accommodate them yet. After I adjusted the API calls in the script, it ran, but I'm not seeing any results. Could you please take a look and help me figure out what's going on?

Thanks so much for your time!

https://github.com/NVIDIA/TransformerEngine/blame/6f4310d700f7445fd12d524c645b1e72fb8886f7/benchmarks/attention/benchmark_attention.py#L200C17-L200C17

output after fixed:

       cuDNN fwd+bwd (ms)  flash-attn fwd+bwd (ms)  cuDNN vs flash speedup
test_0              0.0000                        0                       0
test_1              0.0000                        0                       0
test_2              3.0056                        0                       0
test_3              0.0000                        0                       0

Contributor guide

Open the contributing guide

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.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.