Example of using MeshWorker::Assembler::MGMatrixLocalBlocksToGlobalBlocks?

18 views
Skip to first unread message

Fred

unread,
Feb 17, 2017, 11:14:24 PM2/17/17
to deal.II User Group
Hello,

I currently have a few codes written in the meshworker framework that solve systems of PDE's. (Elasticity, Laplace/Poisson, Stokes) I am interested in experimenting with Multigrid preconditioner for these.  I have been unable to get this to work using the MeshWorker::Assembler::MGMatrixLocalBlocksToGlobalBlocks function.  I am interested in seeing an example of a code where this strategy is used.  If there is any interest in seeing my code, I will be happy to produce a minimal non-working example.  For now, the compilation error that I get from clang is as follows (g++ is similar).


git/dealiiBase/dealiiInstall/include/deal.II/lac/matrix_block.h:932:10: error:
      no viable conversion from returned value of type 'const
      MGLevelObject<dealii::SparseMatrix<double>>' to function return type 'const
      MGLevelObject<MatrixBlock<dealii::SparseMatrix<double> >>'
  return *matrices.read<const MGLevelObject<MatrixType>* >(i);

The line that triggers it is the following

dealii::MeshWorker::integration_loop<dim,dim>
  ( dof_handler.begin_mg(),
    dof_handler.end_mg(),
    dof_Info,
    info_box,
    LDGintegrator,
    assembleSystem);

LDGintegrator is my custom integrator of a type inheriated from  MeshWorker::LocalIntegrator<dim>;
and assembleSystem is of type

MeshWorker
  ::Assembler
  ::MGMatrixLocalBlocksToGlobalBlock< SparseMatrix<double>,
                                                            double>;


Any help would be greatly appreciated.  I would be happy to contribute any changes required to get this to work.

Kent






Wolfgang Bangerth

unread,
Feb 17, 2017, 11:51:07 PM2/17/17
to dea...@googlegroups.com
Can you open an issue on the github forum at
https://github.com/dealii/dealii/issues
It would be great to have a minimal example that shows the problem.

Best
W.

--
------------------------------------------------------------------------
Wolfgang Bangerth email: bang...@colostate.edu
www: http://www.math.colostate.edu/~bangerth/

Fred

unread,
Feb 18, 2017, 12:13:34 PM2/18/17
to deal.II User Group, bang...@colostate.edu
Hi,

I am preparing an example, and now there is better than even chance I figured it out.

I will put up the example on github only if I can't figure it out.

I really appreciate your help!

Fred

unread,
Feb 20, 2017, 11:56:23 AM2/20/17
to deal.II User Group, bang...@colostate.edu
Unfortunately, I wasn't able to figure it out.  I have posted an Issue on github.

Kent
Reply all
Reply to author
Forward
0 new messages