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
Add failing test for multiple ranks and threads #16336
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.
This is a good test. I checked #16340 with it. I would suggest to run it only with 2 MPI rank because I think that exposes the bug already. We should try not to use more resources than necessary, because we do MPI + threads, causing a lot of load on the tester.
mpi_initlog(); | ||
|
||
do_test<2,double>(3,3); | ||
do_test<3,double>(3,3); |
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.
Does the 3D test need to be such big? I would hope that 2 refinements and a polynomial degree 2 should suffice (the 2D variant is ok).
You need to run the indent script on this test. |
I suggest waiting with merging this test until #16340 is merged, to be sure things are fine. |
This is fine for me. Thank you for the quick fix :) |
/rebuild |
Could you please rebase this PR to pass the test? |
ff982e1
to
83cece4
Compare
9bc81ef
to
7baadcf
Compare
@kronbichler @peterrum
MatrixFree::reinit()
fails with multiple ranks and threading enabled. I traced this down to the update of values on inner faces. The error readsUsing only one thread (
Utilities::MPI::MPI_InitFinalize mpi_init(argc, argv, 1)
) works as expected. Running with one rank and threading enabled works as well.