I'm working on a hypothesis that some dG degrees of freedom can be
constrained without compromising the stability for e.g. advection
diffusion equations, so I want a mesh that has some dG and some cG
elements, which is quite unusual.
I see. If the number of constraints is small, you can still go the route via
ConstraintMatrix::condense
after assembling the linear system without considering constraints as a backup
plan. It's just going to be expensive if you have many complicated
constraints.
To my knowledge, the constraint matrix enters data into a sparse matrix
much faster than if you do it by hand. So, John, I think that's a good
idea. Let us know if you made it work!
Guido
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
