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

Understand OpenSBLI performance differences #8

Open
aturner-epcc opened this issue May 31, 2018 · 1 comment
Open

Understand OpenSBLI performance differences #8

aturner-epcc opened this issue May 31, 2018 · 1 comment
Labels
analysis Analysis required data needed Benchmarking data required

Comments

@aturner-epcc
Copy link
Contributor

The Thomas results for OpenSBLI do not seem to follow the performance trends we expect to see. Why is this? Do we see an improvement for Tesseract in the same way for Peta4-Skylake with the additional memory channels?

@aturner-epcc aturner-epcc added data needed Benchmarking data required analysis Analysis required labels May 31, 2018
@owainkenwayucl
Copy link
Contributor

I think at least some of this is related to compiler options on Thomas. I've done some experimentation using the compiler options for CSD3-Skylake on Thomas and other UCL systems and they dramatically improve the performance.

I've put Makefile.ucl files in the repo with these options for OpenSBLI and updated the Thomas build instructions.

I'll commit back results as they come through the scheduler.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
analysis Analysis required data needed Benchmarking data required
Projects
None yet
Development

No branches or pull requests

2 participants