[petsc-dev] DMPlex overlap redistribution (was: Build error with zlib)

[Adrian Croucher]

hi Matt,

On 20/01/24 6:03 am, Matthew Knepley wrote:
>
> I put in a test for redistribution with overlap, Plex test ex12, test 
> 7. You can run it using
>
>   make -f ./gmakefile test search="dm_impls_plex_tests-ex12_7"
>
> and see the output in src/dm/impls/plex/tests/output/ex12_7.out. The 
> second DMPlexDistribute() call
> is made on a mesh with overlap.
>
> If this is not a solution, can we make a small test that goes in ex12 
> that shows what is wrong?

I just got around to testing this with PETSc 3.20.5.

I am still getting problems with redistribution of MINC (dual porosity) 
meshes. The function DMPlexCheckPointSF() is now raising errors like this:

[0]PETSC ERROR: Petsc has generated inconsistent data
[0]PETSC ERROR: Point SF contains 25 which is a cell

Looking in the PETSc code it appears some extra checks were added about 
a year ago (commit 7726db96) including the bit commented "Check there 
are no cells in interface", which is what is raising the error.

My understanding of the point SF is a bit rusty now but I would have 
thought that for finite volume meshes with overlap 1 (which is what I 
have), there should be cells in the interface?

Cheers, Adrian

-- 
Dr Adrian Croucher
Senior Research Fellow
Department of Engineering Science
Waipapa Taumata Rau / University of Auckland, New Zealand
email:a.croucher at auckland.ac.nz
tel: +64 (0)9 923 4611
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20240301/fedaa472/attachment.html>