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

Collect histogram of process loads (modeled and raw) in LB data reduction #1382

Open
PhilMiller opened this issue Apr 14, 2021 · 3 comments · May be fixed by #1873
Open

Collect histogram of process loads (modeled and raw) in LB data reduction #1382

PhilMiller opened this issue Apr 14, 2021 · 3 comments · May be fixed by #1873
Assignees
Labels

Comments

@PhilMiller
Copy link
Member

PhilMiller commented Apr 14, 2021

There's already an implementation of an approximate histogram for the diagnostics. Using that in the process stats reduction would allow quantile estimation, with potential uses in LB strategies.

@PhilMiller
Copy link
Member Author

One possible use would be to iteratively focus refinement-based load balancing on the most-overloaded processes, as judged by their place in the ranking, rather than their absolute degree of overload which is the only thing they can compute with available stats right now.

@lifflander
Copy link
Collaborator

@ppebay Are you still planning on doing this?

@lifflander lifflander changed the title Collect histogram of process loads (total and subphase) in LB stats reduction Collect histogram of process loads in LB stats reduction May 12, 2022
@lifflander lifflander changed the title Collect histogram of process loads in LB stats reduction Collect histogram of process loads (modeled and raw) in LB data reduction May 12, 2022
@stmcgovern stmcgovern linked a pull request Jul 19, 2022 that will close this issue
@stmcgovern stmcgovern linked a pull request Jul 19, 2022 that will close this issue
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging a pull request may close this issue.

4 participants