Hello, Thank you for the reply. I totally get what you are saying about the block structure existing purely as a tool for solvers such as Schur complements. However, after some thought, I wouldn't know how to do some of the "tricks" I'm performing when it comes to matrix assembly outside of a block context.
Before I said that my system is a 4 by 4 block system (some blocks having dim components, others being scalar). However, only one block changes with time. Thus I can save lots of time in assembly using system_matrix.block(0,0) = 0; Followed by only assembling the corresponding block and skipping all other calculations with an if statement. I think the assembly function would still work if system_matrix were not a block matrix, but I haven't been able to discover an alternative for setting just one part of the matrix to zero. Unfortunately, forming only part of the matrix is such a massive speedup that it would be painful to have to drop that optimization. Could you suggest an alternative? Regards, Kyle On Monday, November 29, 2021 at 11:46:32 PM UTC-5 Wolfgang Bangerth wrote: > > > I wasn't sure if I should start a new thread as this is a very related > and > > basic question. Before I got the impression that switching from an exact > LU > > (UMFAPCK) to an ILU would be a good first try when UMFPACK started too > much > > memory for my problem. However, it appears that the SparseILU class does > not > > play nicely with BlockSparseMatrix. > > That is correct. But if you want to just build a preconditioner for the > entire > matrix, then there is also no reason to substructure the matrix into > individual blocks -- you might as well work with one matrix and one vector > that correspond to all variables at once. > > Block matrices/vectors are a tool to build preconditioners that work on > specific blocks of a matrix. If that's not your goal, there is no need to > use > this tool. > > Best > W. > > -- > ------------------------------------------------------------------------ > Wolfgang Bangerth email: [email protected] > www: http://www.math.colostate.edu/~bangerth/ > > -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/dealii/41659ef6-278b-40ce-92dd-6628ef1f2fc3n%40googlegroups.com.
