Sean Miller

@jhux2 I'm not sure if you guys are still tracking MueLu_DriverDiagonalModifications_MPI_1 down, but it can be fixed by commenting out MueLu_FilteredAFactory_def.hpp:240. The const cast of the host view (getLocalRowView on...

@lucbv Do you happen to know who can review this? The cmake/naming changes are somewhat invasive.

@csiefer2 Feel free to make modifications, I wasn't sure what to do about documentation, or updating any of the ATDM builds if that was necessary, so I am expecting some...

@csiefer2 Thanks for looking into this. I was confused by the requirement for the kokkos::fence()'s I had added. I had originally thought that the deep_copy operations had fences internally, but...

@cgcgcg Here are the changes we spoke about previously. It should be useful for any application that uses the classical uncoupled aggregation AMG scheme, where the symbolic data never changes,...

> @seanofthemillers Where is this actually reusing the coloring? Ha, I missed a copy/paste somewhere. I'll fix that it a moment.

@cgcgcg You mentioned that I should change the name from "aggregation: reuse coloring" = ON/OFF. Would "reuse: type" = "coloring" be preferable?

@jhux2 I would be happy to, but I'm curious, is there a way to check if these options are being used? As it stands, the solver will work fine if...

@cgcgcg @jhux2 I have addressed your comments (I think) and have added a new test, though it is a little limited. I have used a tracing API to confirm that...

@cgcgcg Ah, this might just be me misunderstanding what tentative P is. For the problem I'm looking at, I am assuming that the values stored in matrix A are changing...