Closed MedAdh42 closed 1 year ago
@dkachuma @francoishamon can you take a look?
I cannot reproduce the issue on develop
. I believe that this issue was resolved by #2161. I can reproduce the issue with 16089ad4b but when I switch to 8950803bc, the model is able to initialise and start time stepping.
@jeannepellerin @MedAdh42 Are you able to run this with a more recent version?
Hi @dkachuma, Unfortunately, the last version we have on P2 is dated Nov 4th, so before the resolving issue you mentionned. Is it possible to deploy a new version ?
Hi @dkachuma, Unfortunately, the last version we have on P2 is dated Nov 4th, so before the resolving issue you mentionned. Is it possible to deploy a new version ?
Hello @margauxraguenel, I've just compiled the latest version of develop repo on P2. Please check if you can get access to it:
/workrd/SCR/GEOSX/install/gcc8/GEOSX-9d7cd0d/bin/geosx
Thank you for your responses, the issue is solved. I'll go ahead and close it.
Describe the bug I'm testing different meshes of the same geological model with the same simulation parameters running on 8 MPI ranks. The reference model was successfully run by Margaux Raguenel. I have 16 different meshes and at least 3 of them fail, GEOSX throws an error located at : WellElementRegion.cpp:75.
Two examples (fail and success) are given: GEOSX_Perforation_bug.zip
Expected behavior The perforation is at the same geometrical location within the bounds of the domain for all the meshes. The perforated cells should be found for all meshes.
Platform :
Additional context We suspect that the issue is related to MPI partitions since it successfully run in serial @jeannepellerin .