You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At the moment, the handling of refinement of ghosted meshes is a mess. It should be easy to optionally include ghost cells in a refined mesh without re-partitioning by using compute_destination_ranks(...) in graph/partitioners.cpp.
This is compounded by transfer_facet_meshtag(...) and transfer_cells_meshtag(...) no supporting ghosted meshes, or re-partitioned refined meshes. Both should be relatively easy to support. This leads to unexpected 'user experiences'.
It could help a lot if mesh::create_mesh returned the destination ranks of the input cells.
The text was updated successfully, but these errors were encountered:
At the moment, the handling of refinement of ghosted meshes is a mess. It should be easy to optionally include ghost cells in a refined mesh without re-partitioning by using
compute_destination_ranks(...)
ingraph/partitioners.cpp
.This is compounded by
transfer_facet_meshtag(...)
andtransfer_cells_meshtag(...)
no supporting ghosted meshes, or re-partitioned refined meshes. Both should be relatively easy to support. This leads to unexpected 'user experiences'.It could help a lot if
mesh::create_mesh
returned the destination ranks of the input cells.The text was updated successfully, but these errors were encountered: