I'm looking into the implicit neighbor dof stuff in the sparsity pattern generator.... and I have a few observations.
1. Do you think it would be possible to change the default for implicit_neighbor_dofs such that if you have _any_ discontinuous variables you switch it to true? Right now you have to have _all_ discontinuous variables for it to default to true.... and that messes up mixed formulations where you have some DG and some CG going on. I can't think of a time when you would have discontinuous variables that you are solving for and you wouldn't want implicit neighbor dofs on. Maybe if you're just doing an L2 projection of some data into a discontinuous function space? 2. Is there a reason we have to do implicit_neighbor_dofs for all variables or none at all? Shouldn't it be possible to generate a per variable sparsity pattern so that only the ones that are discontinuous get the extra non-zeros? I obviously haven't looked closely enough at the code... but I thought maybe someone could derail that train of thought for me before I even get started. Thanks for any insight! Derek ------------------------------------------------------------------------------ This SF.net email is sponsored by Make an app they can't live without Enter the BlackBerry Developer Challenge http://p.sf.net/sfu/RIM-dev2dev _______________________________________________ Libmesh-devel mailing list Libmesh-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-devel
