[dune-pdelab] artifact in DiscreteGridfunctionGradient in parallel run

[Michael Wenske]

I'm sorry I should have provided that right away!

I wrote a short MWE. All it does is create a vector into which I
interpolate a sine function.
Then I use the gradientAdapter and write the gradient it to vtk. The
field itself is written without any problem,
but the gradient field shows a tiny, but relevant disturbance at the
processor boundaries.

The MWE is in the form of a small dune-module.

*build with:*

./dune-common/bin/dunecontrol --only=gradReproducer all

*run with: *

mpirun -n 2 ./gradReproducer 

Reproducer dune module and main.cc in appendix.

Thanks in advance,

Michael



On 01.08.19 14:28, Jö Fahlke wrote:
> Hi Michael,
>
> could you point us to your reproducer, or at least the output of the
> reproducer?
>
> Regards,
> Jö.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dune-project.org/pipermail/dune-pdelab/attachments/20190801/55017b27/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: gradReproducer.zip
Type: application/zip
Size: 8623 bytes
Desc: not available
URL: <https://lists.dune-project.org/pipermail/dune-pdelab/attachments/20190801/55017b27/attachment.zip>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: gradReproducer.cc
Type: text/x-c++src
Size: 3763 bytes
Desc: not available
URL: <https://lists.dune-project.org/pipermail/dune-pdelab/attachments/20190801/55017b27/attachment.cc>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: OpenPGP digital signature
URL: <https://lists.dune-project.org/pipermail/dune-pdelab/attachments/20190801/55017b27/attachment.sig>