Derek Gaston
Derek Gaston
Gotcha - thanks for the explanation! It sounds reasonable to switch it.
I would accept a PR for this!
Is it not enough to just do a `module list` and `env` and the beginning of every test? That should capture it well enough.
Good luck with that. This one might come down to a "procedure" for @milljm . If he modifies the environment of the test boxes he may need to make note...
I will think this should be put on hold until someone wants to do it using the GhostingFunctor stuff. That is the right way to do it. @roystognr even built...
BTW: `need_full_sparsity_pattern` will actually use far MORE RAM! The SparsityPattern object itself takes a pretty huge amount of memory (and a lot of time) to build. Rerun your first stuff...
@fdkong Yes... the gometric search system will feed the GhostingFunctors the right information. Then the sparsity pattern should be based on the GhostingFunctors.
@fdkong did you delete a message or something? I got an email about another message but I don't see it. It looks like you were questioning about why setting `need_full_sparsity_pattern`...
In particular: check this: https://github.com/libMesh/libmesh/blob/master/src/base/sparsity_pattern.C#L346