-
Notifications
You must be signed in to change notification settings - Fork 29.8k
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
benchmark: update iterations in benchmark/crypto/aes-gcm-throughput.js #50929
Conversation
I don't think that worth changing from |
Hi Vinicius, The rational of this change is not to trigger JIT, but make the crypto code dominate the execution phase instead of fwrite. When the iteration was 500, most execution time was occupied by fwrite. When iterations was increased to 2500, most execution time would be occupied by crypto. To make the scoring reasonable, we need to increase the iterations. After some experiments I found 2500 is a proper value to avoid long execution time. I attached the perf log of 500 and 2500 and please see if the change made sense. Thanks! |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In this case, make sense, thanks for the PR.
Landed in 29af78e |
Fixes: #50571 PR-URL: #50929 Refs: #50571 Reviewed-By: Vinícius Lourenço Claro Cardoso <[email protected]> Reviewed-By: James M Snell <[email protected]>
Fixes: #50571 PR-URL: #50929 Refs: #50571 Reviewed-By: Vinícius Lourenço Claro Cardoso <[email protected]> Reviewed-By: James M Snell <[email protected]>
Refs: #50571
Below is score improvement from 500 to 2500, measured on Cascake server.
Improvement can be a bit more if changing to 5000 but the running time will be notable longer than appropriate duration.
To make it clear, I will submit more PRs one by one to address each case in crypto, and tracked by 50571 issue.