Dear all,
I've got matrix-free multigrid solver for Stokes problem. The main
bottleneck is solution of coarse problem, so I tried to assemble the
regular sparse matrix and use direct solver. Since the coarse problem is
(relatively) small, I used vmults by unit vector to obtain columns of the
matrix. This is my code:
setup_matrix(); // generates sparsity patters and reinits matrices
LinearAlgebra::distributed::BlockVector<LevelNumber> dst;
LinearAlgebra::distributed::BlockVector<LevelNumber> src;
src.reinit(2);
for (unsigned int b = 0; b < 2; ++b)
stokes_matrix.get_matrix_free()->initialize_dof_vector(
src.block(b), b);
src.collect_sizes();
src =0;
dst.reinit(2);
dst.block(0).reinit(owned_dofs_u, relevant_dofs_u, MPI_COMM_WORLD);
dst.block(1).reinit(owned_dofs_p, relevant_dofs_p, MPI_COMM_WORLD);
dst.collect_sizes();
for(types::global_dof_index i =0; i< owned_dofs_u.size(); ++i){
src=0;
dst=0;
if(owned_dofs_u.is_element(i) )
src.block(0)(i)=1;
src.compress(VectorOperation::insert);
stokes_matrix.vmult(dst, src);
dst.update_ghost_values();
for(IndexSet::ElementIterator index_iter =owned_dofs_u.begin();
index_iter != owned_dofs_u.end();
++index_iter){
if(dst.block(0)(*index_iter)!=0 )
block_A->set(*index_iter,i, dst.block(0)(*index_iter) );
}
}
block_A->compress(VectorOperation::unknown);
Without constrains the matrix matches the matrix-free operator, but with
constrains present it does not. What is the proper way to assemble the
matrix with vmult?
Best,
Michał Wichrowski
--
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/c1bf23f7-a8fb-41cb-b24c-877afc1b16b1%40googlegroups.com.
