[deal.II] Re: Adding an independent DoF to a Trilinos system

2018-03-22 Thread Ben Shields
I did some digging into the Trilinos reinit_matrix, and I've found that all of my lost time is coming from the operation graph->FillComplete(input_col_map, input_row_map) This step is accounting for as much as 20 seconds per call for a 1x200,000 matrix on 4 processors. I also found that this

[deal.II] Re: Adding an independent DoF to a Trilinos system

2018-03-20 Thread Ben Shields
Thank you so much for your help Wolfgang. I'll keep looking into that line which is still causing me trouble, but hopefully this will alleviate the biggest chunk of the bottleneck. -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.go

[deal.II] Re: Adding an independent DoF to a Trilinos system

2018-03-20 Thread Ben Shields
> > Can you narrow down where the time is lost? These two lines added_sp.block(0, 1).copy_from(col); added_sp.block(1, 0).copy_from(row); account for about 60% of the bottleneck (considering only the runtime for the piece of code in the original post). The other 40% of the bottleneck is f

[deal.II] Re: Adding an independent DoF to a Trilinos system

2018-03-19 Thread Ben Shields
It is setting up the matrices which is the bottleneck, specifically setting up the off-diagonal blocks of the matrix. For context, with a moderately sized system of 200,000 dofs, setting up the main NxN sparsity pattern using DoFTools::make_sparsity_pattern and reinitializing the matrix using t

[deal.II] Adding an independent DoF to a Trilinos system

2018-03-15 Thread Ben Shields
Hello all, I'm trying to add a single DoF to my Trilinos system (not associated with the FE system, similar to the question asked here ). I followed the advice there, and set it up as a block system, where block (0,0) is the usua

[deal.II] Re: Trilinos SparseMatrix mmult bug

2016-09-29 Thread Ben Shields
So this turned out to be a bug in my own code after I made the original change in trilinos_sparse_matrix.cc. When I was constructing the sparsity patterns for the matrices using DoFTools::make_sparsity_pattern, there is a flag I had set keep_constrained_dofs = false. Setting this flag to true ha

[deal.II] Re: Trilinos SparseMatrix mmult bug

2016-09-23 Thread Ben Shields
CENSE at * the top level of the deal.II distribution. * * - * * Authors: Martin Kronbichler, Uppsala University, * Wolfgang Bangerth, Texas A&M University, * Timo Heister, University of Goettingen, 2008-2011 *

[deal.II] Re: Trilinos SparseMatrix mmult bug

2016-09-22 Thread Ben Shields
I've just updated to 8.4.2 and repeated the change to trilinos_sparse_matrix.cc, and the matrices appear the same as my previous post. So I've at least ruled out that it's not an issue specific to my version number. -- The deal.II project is located at http://www.dealii.org/ For mailing list/f

[deal.II] Re: Trilinos SparseMatrix mmult bug

2016-09-22 Thread Ben Shields
Hey there, thanks for the response. I made the recommended change to trilinos_sparse_matrix.cc, and the corresponding test code returns the appropriate output. However, I'm still getting an issue in my particular case. It's improved in that the resulting matrix is now square and the dimensions

[deal.II] Trilinos SparseMatrix mmult bug

2016-09-21 Thread Ben Shields
ult using 2 processors in parallel. Any help would be greatly appreciated. Cheers, Ben Shields -- 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 subscr