[dune-pdelab] Error in testgeneo.cc

[Aswin vs]

Dear Linus Seelinger,

  Onceagain thank you so much for your great help. Still I have some
convergence issue of the linear solver in the same example by replacing
ALUgrid.
The inserted lines are here. Here I am constructing the overlapping grid
using the ALUgrid.  I greatly appreciate if you provide any suggestions in
this regard.

typedef Dune::ALUGrid<dim,dim, Dune::cube ,Dune::nonconforming> GM;
  Dune::StructuredGridFactory<GM> factory;
Dune::FieldVector<double,2> lowerLeft(0.0);
     Dune::FieldVector<double,2> upperRight(1.0);
     std::array<unsigned int,dim> N1 = {100, 100};
auto grid = factory.createCubeGrid(lowerLeft,upperRight,N1);
grid->loadBalance();

Thank you..
-Aswin VS




On Mon, Sep 27, 2021 at 6:22 PM Linus Seelinger <
linus.seelinger at iwr.uni-heidelberg.de> wrote:

> Hi Aswin,
>
> I assume you are running the test binary sequentially? The GenEO coarse
> space is only valid for a domain decomposition with overlaps (sequential
> run => no grid overlap => empty right hand side in GenEO eigenproblem =>
> arpack failure).
>
> You can look up the definition of tests in the corresponding
> CMakeLists.txt file. There you'll see that this test is intended to run on
> 2 MPI ranks, so
> mpirun -n 2 ./testgeneo
> should do the trick.
>
> Best,
>
> Linus
> Am Montag, 27. September 2021, 13:40:18 CEST schrieb Aswin vs:
> > When I try to run testgeneo.cc from the pdelab/text directory, I receive
> > the following error message.
> > Could someone please advise me how to solve this issue?
> >
> > Arpack error in Aupp.
> > -> Starting vector is zero.
> > Arpack error in FindArnoldiBasis.
> > -> Could not find an Arnoldi basis.
> > Process 0 picked 0 eigenvectors
> > Matrix setup
> > Global basis size B=0
> > Matrix setup finished: M=3.2284e-05
> > === Dune::CGSolver
> >  Iter          Defect            Rate
> >     0      9.97497e+06
> >     1      1.29537e-08      1.29862e-15
> > Coarse time CT=7.905e-06
> > Coarse time per apply CTA=7.905e-06
> > Arpack error in AREigenProblem.
> > -> 'n' must be greater than one.
> > Arpack error in DefineParameter.
> > -> Some parameters were not correctly defined.
> > Arpack error in TakeStep.
> > -> DefineParameters must be called prior to this function.
> > Arpack error in FindArnoldiBasis.
> > -> Could not find an Arnoldi basis.
> > [user-OptiPlex-7080:41995] *** Process received signal ***
> > [user-OptiPlex-7080:41995] Signal: Segmentation fault (11)
> >
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dune-project.org/pipermail/dune-pdelab/attachments/20210929/0fccb7ee/attachment.htm>