search

fdrmrc / Polydeal

C++ implementation of Polygonal Discontinuous Galerkin method within the deal.II Finite Element library.
https://fdrmrc.github.io/Polydeal/
Other
0 stars 0 forks source link

Use only Trilinos sparsity pattern in parallel #114

Closed fdrmrc closed 6 months ago

fdrmrc commented 6 months ago

Fixes https://github.com/fdrmrc/Polydeal/pull/113#issuecomment-2126968166, only 7001cd130b26fa38dbeb7d87f122659915e5f01f is relevant, the other commits are related to tests.

This PR improves the creation of the sparsity pattern in the distributed case by using the Trilinos version TrilinosWrappers::SparsityPattern. If not running with MPI, the standard DynamicSparsityPattern class is used. The relevant change is:

https://github.com/fdrmrc/Polydeal/blob/79df2257e5a053a86d37b99fb6076330206f5fc9/source/agglomeration_handler.cc#L679-L687

fdrmrc commented 6 months ago

The same reasoning above has to be applied also to the post-processing routine, where we interpolate onto the finer grid.

fdrmrc commented 6 months ago

Tests against latest deal.II master are all passing, so I'm merging.