Dear Dr. Arndt,
Thank you very much for your help. The problem was resolved applying your
comment. I was setting the indicators running over locally owned cells not
all the cells.
Best regards,
Hamed
On Saturday, February 17, 2018 at 11:20:32 AM UTC-6, Daniel Arndt wrote:
>
> Hamed,
>
> for c
Hamed,
for collecting the periodic boundary pairs only the coarsest level is
important.
In particular, you have to make sure that you set the boundary indicators
correctly for all cells on this level.
For the coarsest level every MPI process knows about all the cells since
the mesh is the same