-
Notifications
You must be signed in to change notification settings - Fork 37
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
Fix bug in static refinement #1071
Conversation
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.
Thanks for the catch! To prevent this kind of thing from happening again... maybe we should have a regression test that uses SMR?
Agreed, we need to add a couple tests that use multi-tree forests. We have at least one that uses SMR, but the bug does not manifest if the base grid only requires one tree. |
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.
👍 for fixes and 👍 for tests
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.
LGTM
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.
LGTM (plus optional comments on the python stuff)
The updated gold files are in #1073, right?
+ "/tst/regression/gold_standard/sparse_twotree.out0.final.phdf", | ||
], | ||
one=True, | ||
tol=1e-12, |
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.
Any reason for the "loose" tolerance?
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.
I think I just copied this tolerance from the other checks. Isn't 1.e-12 what we set for most tests? I don't think this test should be bitwise exact between machines.
PR Summary
The re-definition of
Mesh::root_level
introduced in #1053 did not get correctly propagated through the static refinement related code (which operates on logical locations relative to the legacy tree). This PR fixes that issue, which had cropped up in Artemis.PR Checklist