Reduce MPI communication overhead

[JTHesse]

What are you trying to do and how would you want to do it differently?

Currently, the peridynamic calculations are quite efficient but the MPI communication needs to be improved.

Is it something you currently cannot do? Is this related to an issue/problem?

No response

Implementation details

We need to determine the bottleneck and benchmark it.

Would you be willing to submit a PR?

Yes