test_periodic_boundary_computation.py is empty

[mscroggs]

Every line in python/test/unit/mesh/test_periodic_boundary_computation.py is commented out.

We should either remove this file, or uncomment these tests.

[jorgensd]

This has not been supported for a long time (and is partially the reason why I made dolfinx_mpc. However, for the special case that the mesh nodes/facets align on each side of the periodic boundary, it would be neat to have a more efficient code than what I got (i.e. something that just alters the dofmap).

[garth-wells]

This has not been supported for a long time (and is partially the reason why I made dolfinx_mpc. However, for the special case that the mesh nodes/facets align on each side of the periodic boundary, it would be neat to have a more efficient code than what I got (i.e. something that just alters the dofmap).

See #18. Is there a reason why the dolfinx_mpc approach will be noticeably slower than a more integrated approach when facets match? Can dolfinx_mpc handle periodic DG?

[jorgensd]

Is there a reason why the dolfinx_mpc approach will be noticeably slower than a more integrated approach when facets match?

DOLFINx_MPC does not alter the dofmap, as this is not possible in most cases (only with grid alignment and periodic coefficient=1 would this be the case, i.e. $u(dof(x_i)) = u(dof(x_j))$.

Can dolfinx_mpc handle periodic DG?

I have not tested it with periodic DG, but I think it should be possible by using: https://github.com/jorgensd/dolfinx_mpc/blob/main/python/dolfinx_mpc/multipointconstraint.py#L129-L145