On 1/17/24 14:40, Alex Quinlan wrote:
If I do this, then it seems that I'd end up with two conflicting
partitions. If I have n_mpi_processes=8 , for example, this approach
would assign ~1/8th of the solid mesh to process 1 AND ~1/8th of the
shell mesh to process 1. This seems like a problem.
No, this is exactly what you want. That's because you will have phases
in your program when you work on the solid mesh, and you want all 8
processes to participate in that. And then you have a phase where you
going to do something on the surface mesh and, again, you want to have
all 8 processes work on that.
It is *so* much harder to write software in which different processes do
different things at the same time. Don't go there.
As you pointed out, I will have an issue with the off-diagonal matrices
and the coupling between the two triangulations. Is it possible for me
to build up all of the sparsity patterns and matrices /and then /perform
the partitioning?
You could of course, but then what's the point of using multiple
processors if you do the bulk of the work on only one process?
Best
W.
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/dealii/2f0baf30-2db6-4863-b4c0-586cc36fc834%40colostate.edu.
