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

UCP/WIREUP: Consider local distance during slow lanes dropping #9512

Conversation

ivankochin
Copy link
Contributor

What

Make lanes filtering function aware of local distance.

Why ?

Remote BW value already capped by distance during the comparison, but local BW doesn't consider distance. That can cause reconfiguration errors because the peers can calculate different BW numbers for the same lane.

@ivankochin ivankochin self-assigned this Nov 29, 2023
Copy link
Contributor

@pentschev pentschev left a comment

Choose a reason for hiding this comment

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

I can confirm this resolves the issue I was experiencing on my end. Thanks @ivankochin !

bw_local = ucp_tl_iface_bandwidth(worker->context,
&wiface->attr.bandwidth);
} else {
bw_local = ucp_wireup_iface_bw_distance(wiface);
Copy link
Contributor

Choose a reason for hiding this comment

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

pls add a short comment what it fixes
👍 otherwise

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Done.

yosefe
yosefe previously approved these changes Nov 30, 2023
@ivankochin ivankochin force-pushed the ucp/wireup/consider-local-distance-duing-slow-lanes-dropping branch from 47542e4 to 350d795 Compare November 30, 2023 11:38
@ivankochin ivankochin merged commit 8120597 into openucx:master Dec 1, 2023
115 checks passed
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

Successfully merging this pull request may close these issues.

4 participants