Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[Bug] CK GeMM correctness bug #199

Open
chhwang opened this issue Jan 4, 2024 · 1 comment
Open

[Bug] CK GeMM correctness bug #199

chhwang opened this issue Jan 4, 2024 · 1 comment

Comments

@chhwang
Copy link
Contributor

chhwang commented Jan 4, 2024 •
edited
Loading

Describe the bug
test_matmul_fp32 and test_matmul_fp16_split fails on MI300x.

To Reproduce
Run the unit test.

Expected behavior
max_diff should be lower than the calculated value.

System (please complete the following information):

  • MI300x
  • Single-GPU

Additional context
@Binyang2014
Copy link
Contributor

test_matmul_fp16_split is caused by program logical issue.
When we split the output tensor and input tensor, due to input tensor cannot match the tile size, it will be padded input another shape. Then the results for each part will overwrite the same output region.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@chhwang @Binyang2014