p-refinement of mesh geometry#4205
Conversation
…o jhale/gdim-tdim-mesh-create
| basix::FiniteElement<T> b_element = basix::create_element<T>( | ||
| basix::element::family::P, | ||
| mesh::cell_type_to_basix_type(new_cmap.cell_shape()), new_cmap.degree(), | ||
| new_cmap.variant(), basix::element::dpc_variant::unset, false); |
There was a problem hiding this comment.
I was thinking about this on the way home, and could we potentially here make a DG geometry if we would like to, i.e. have an input option for setting discontinuous to True here?
See for instance the tests in IO4Dolfinx that goes from 1->3->1 and 1->2->1: |
|
It's useful for IO as well, as it would allow you to read/write point data on higher order grids (and then linearize the grid after the fact). Additionally, as the meshes share topology, it is easy to do interpolation between the two grids. |
2 participants
Adds a single free function that can p-refine (and possibly coarsen?) an existing mesh.
Testing a bit limited: P1 to P2 and P1 to P1. Ideas welcome.