-
Notifications
You must be signed in to change notification settings - Fork 49
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
Enable Mi300 Roofline #285
Comments
Happy to help. Let me know what you need. |
Add me to this as well. @coleramos425 you want to set something up ? |
I'll add you to the meeting @skyreflectedinmirrors |
hi @coleramos425 is mi300 roofline enabled in omniperf? get similar request, thanks! |
@lizamd we're still developing roofline support. If you ping me on Teams I can provide more detail. |
cool, thanks for the quick response, will ping you! |
Hi @coleramos425, I am also trying to use roofline for bench-marking llama.cpp. Any workarounds or temporary solutions to get it to work would be of great help. Thanks in advance!! |
Update 8/16Just received a beta version of microbenchmark from Rene. @cfallows-amd will be working to validate that work while myself and others work on adjustments to the arithmetic intensity calculation |
Is your feature request related to a problem? Please describe.
Both national labs, internal customers, and others are asking for a Mi300 based roofline analysis. In v2.0 release this was temporarily disabled.
The microbenchmark has been adjusted with help from @rwvo . However, I still want to stress test the arithmetic intensity calculation and put the model through the wringer with some test applications.
Describe the solution you'd like
The text was updated successfully, but these errors were encountered: