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

scripts for performance tracking system #825

Merged
merged 13 commits into from
Oct 26, 2024
Merged

Conversation

jzuniga-amd
Copy link
Collaborator

This PR adds new scripts to generate internal performance reports to track the optimization efforts (see the README file for more details).

@jzuniga-amd jzuniga-amd added the noOptimizations Disable optimized kernels for small sizes for some routines label Sep 26, 2024
@jzuniga-amd jzuniga-amd marked this pull request as ready for review September 26, 2024 19:51
@amd-jnovotny
Copy link
Contributor

I noticed that we have a docs section on benchmarking. Do you think this information needs to be added to docs too? Or is it something that only developers/contributors would be concerned with?

@jzuniga-amd
Copy link
Collaborator Author

I noticed that we have a docs section on benchmarking. Do you think this information needs to be added to docs too? Or is it something that only developers/contributors would be concerned with?

Correct, this is intended only for internal use (by developers and CI)

@samjwu samjwu added the runPerformance Run CI performance job label Sep 30, 2024
POTRFBATCH tests are run with the given precision and sizes
"""
def potrfBatch_suite(*, suite, precision, sizenormal, sizebatch):
fn = 'potrf_batched'
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Other tests test the strided_batched function. Is there a reason POTRF is testing the batched function?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah, all batch functions in cusolver are strided_batched, but for some reason, potrf is _batched

scripts/perf/rocsolver-perfoptim-suite.py Outdated Show resolved Hide resolved
Copy link
Collaborator

@jmachado-amd jmachado-amd left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The code reads well, and everything looks good to me. Yet, I am trusting CI to validate this one, as I have not tried the scripts myself.

@jzuniga-amd jzuniga-amd merged commit 8841f97 into ROCm:develop Oct 26, 2024
0 of 3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
noOptimizations Disable optimized kernels for small sizes for some routines runPerformance Run CI performance job
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants