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
particle problem with deal.ii 9.2.0-pre but not 9.0.1 #3354
Comments
Could you check if it only happens on 2 MPI processes when one process has all the particles (no matter how many) and the other has none? This looks like a problem with the particle transfer during mesh refinement, and in particular something about how much data each process expects to get. |
I looked at the 2 particle case, and they both where in a different file/on a different processor. When I made an ascii file with two points on the same location and ran it again, it produced the same error. So it indeed seems to be a problem on a mpi process having no particles. |
The difference, btw, is whether we use functionality in ASPECT (when using deal.II 9.0) or in deal.II where much of this code has moved later and from which ASPECT takes it when available in deal.II (using 9.2.pre). |
We overhauled the data transfer algorithm with deal.II version 9.1. In its process, we introduced an overstrict assertion while sending variable size data: Every cell was expected to receive variable size data once a corresponding pack/unpack callback has been registered. The logic in the whole send/receive apparatus should be okay though -- it was just this assertion. I guess your problem would vanish if you'll work in release mode. We removed this assertion in dealii/dealii#10215. Hopefully we can fix this with the 9.2 release with dealii/dealii#10216. |
Thanks for letting us know! |
This was fixed in deal.II 9.2.0. deal.II 9.1.0 unfortunately will keep this bug. |
When I run the prm file below (one of the tests but with 1 particle and 1 adaptive mesh refinement) with deal.ii v9.0.1 it runs fine, but when I run it with deal.ii 9.2.0-pre I get the error below.
The problem goes away when I use no mesh refinement and/or 2 or more particles. Any idea's on what the problem could be?
The text was updated successfully, but these errors were encountered: