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.

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

AMD MI210 only got 20% performance of Nvidia A100 of MultiRHSProf - 8rhs #105

Open
Alice1069 opened this issue Aug 8, 2022 · 1 comment
Assignees

Comments

@Alice1069
Copy link

I test latest code on MI210, I got 0.36TOPS,
if i removed #program unroll,I got 0.88TOPS.
compares to nvidia-A100 4.3 Tops, it's only 20%.
considering A100 and Mi210 has similar fp32 ops capability, is it reasonable or still have space to improve mi210 performance?

configs:
Lattice = 64 64 64 20
SIMULATeQCD code: main branch, git commit 2022 Aug 2, 551ccca
-DARCHITECTURE="gfx90a" //for mi210
mpiexec -np 1 --allow-run-as-root ./MultiRHSProf (8rhs)

@clarkedavida
Copy link
Collaborator

Hi Alice, thanks for looking into this, and sorry about the extremely late response.

We have been noticing poor performance on MI250X compared to A100. We are not yet sure what is causing this... We will let you know if we figure out what is causing our performance loss.

@clarkedavida clarkedavida self-assigned this Jan 6, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants