[petsc-users] Problem with interpolating projected field of quadratic order

Matthew Knepley knepley at gmail.com
Sat Jun 18 09:26:40 CDT 2016


On Sat, Jun 18, 2016 at 8:56 AM, Sander Arens <Sander.Arens at ugent.be> wrote:

> Hello all,
>
> Recently I've been trying to run a fem problem using DMPlex, PetscFE, etc.
> using a mesh created with gmsh. I used DMPlexReverseCell to get rid of the
> negative jacobian determinants. However, I noticed another problem: when I
> projected the coordinates on a quadratic fem-field and interpolated them
> they didn't match the interpolated values of the original coordinate field
> anymore.
>
> I attached a simple test to reproduce this. With linear interpolation I
> see no problem, it's only when I start using quadratic interpolation.
>
> I have no idea why this is giving problems. Am I missing something simple
> here or is this a bug?
>

Tests for this already exist


https://bitbucket.org/petsc/petsc/src/996a00f1a1693e2d85044ed633fe1848cf357f2a/src/dm/impls/plex/examples/tests/ex3.c?at=master&fileviewer=file-view-default

with run parameters here:


https://bitbucket.org/petsc/petsc/src/996a00f1a1693e2d85044ed633fe1848cf357f2a/config/builder.py?at=master&fileviewer=file-view-default#builder.py-58

and specifically for P2 triangles


https://bitbucket.org/petsc/petsc/src/996a00f1a1693e2d85044ed633fe1848cf357f2a/config/builder.py?at=master&fileviewer=file-view-default#builder.py-72

1) Does the test run for you?

  ./config/builder2.py check src/dm/impls/plex/examples/tests/ex3.c

2) If so, then could it be that you assume something about the order of
functions?

  Thanks,

     Matt


> Thanks,
> Sander
>



-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20160618/829e8e81/attachment.html>


More information about the petsc-users mailing list