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

Fixed an issue where dymos calling excessive allgathers during setup. #980

Merged
merged 3 commits into from
Sep 6, 2023

Conversation

robfalck
Copy link
Contributor

@robfalck robfalck commented Sep 5, 2023

Summary

dymos had been using allgather to collect units and shapes of variables for linkage calculations, and parameter information for trajectory parameters. These allgathers have been reduced to a single call at the start of the Trajectory configure process to pick up any variable information that was populated in the phases via introspection.

Related Issues

Backwards incompatibilities

None

New Dependencies

None

@@ -530,31 +506,42 @@ def _configure_parameters(self):

def _configure_phase_options_dicts(self):
"""
Called during configure if we are under MPI. Loops over all phases and broadcasts the shape
and units options to all procs for all dymos variables.
Called during configure if we are under MPI. Loops over all phases and populates the
Copy link
Member

Choose a reason for hiding this comment

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

Isn't this called even if you aren't under MPI?

@robfalck robfalck merged commit a129652 into OpenMDAO:master Sep 6, 2023
8 checks passed
@robfalck robfalck deleted the mpi_performance branch August 14, 2024 15:30
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.

Long setup times under MPI
3 participants