Jørgen Schartum Dokken
Jørgen Schartum Dokken
> Is this feature on the horizon? As I haven't gotten any feedback, I haven't moved it to DOLFINx yet. You can use the scifem version, as scifem is on...
> * Paraview fails to recognise 2D data for geometry for 2D meshes (probably needs padding to 3D). Ref the Python implementation we had; I stored a variable gdim which...
We would need to figure out how to distribute the various cell types in a consistent order across all processes. The main issue is that in VTKHDF5, there is no...
It is a good suggestion. The main reason for it being as it is, is that I didn't want to use `@singledispatch` at the time.
2D refinement routines would also be nice (quad, tri) and easier to debug).
> Can you add some gentle background to the PR description? I've added an example to the description of the PR.
@garth-wells any further comments?
@garth-wells I've addressed your comments. Good to merge?
> This is more of a critical bug-fix, so should go in asap. Btw. is the local index now needed? Looks like it will always be 0, 1, 2, ...,...
If we go for VTKHDF, please consider: https://discourse.vtk.org/t/vtkhdf-point-ownership/13913/4 as it is to me unclear how to actually make use of the partitioning offsets in VTKHDF. https://gist.github.com/jorgensd/6b3db748009ac1017929d7e7f3e83dc2 shows how the data-structure...