[Trilinos-Users] trouble anasazi with mkl
SungHwan Choi
sunghwanchoi91 at gmail.com
Sun Jan 6 22:32:11 MST 2013
Hi ,
I have problem on anasazi. my anasazi
Epetra_CrsMatrix* kinetic_matrix ;
kinetic_matrix=new Epetra_CrsMatrix(Copy,Map,0);
Teuchos::RCP< Anasazi::BasicEigenproblem<double,MV,OP> > MyProblem =
Teuchos::rcp( new Anasazi::BasicEigenproblem<double,MV,OP>(matrix, ivec) );
bool ierr = MyProblem->setProblem();
cout << ierr <<endl;
MyProblem->setNEV( 5 );
MyProblem->setHermitian(true);
int verb = Anasazi::Warnings + Anasazi::Errors+ Anasazi::FinalSummary +
Anasazi::TimingDetails;
// Create the parameter list for the eigensolver
Teuchos::ParameterList MyPL;
MyPL.set( "Verbosity", verb );
MyPL.set( "Which", "SM" );
MyPL.set( "Block Size", 8 );
MyPL.set( "Num Blocks", 10 );
MyPL.set( "Maximum Restarts", 100 );
MyPL.set( "Convergence Tolerance", 1.0e-8 );
// Create the Block Krylov Schur solver
// This takes as inputs the eigenvalue problem and the solver parameters
Anasazi::BlockKrylovSchurSolMgr<double,MV,OP>
MyBlockKrylovSchur(MyProblem, MyPL );
// Solve the eigenvalue problem, and save the return code
Anasazi::ReturnType solverreturn = MyBlockKrylovSchur.solve();
// Check return code of the solver: Unconverged, Failed, or OK
switch (solverreturn) {
// UNCONVERGED
case Anasazi::Unconverged:
if (verbose){
cout << "Anasazi::BlockKrylovSchur::solve() did not
converge!" << endl;
}
#ifdef HAVE_MPI
MPI_Finalize();
#endif
return ;
break;
// CONVERGED
case Anasazi::Converged:
if (verbose){
cout << "Anasazi::BlockKrylovSchur::solve() converged!" <<
endl;
}
break;
}
// Get eigensolution struct
Anasazi::Eigensolution<double, Epetra_MultiVector> sol =
MyProblem->getSolution();
and I compiled with -lmkl_intel_lp64 -lmkl_core -lmkl_sequential. I hope
this will going well but I got error message as below
terminate called after throwing an instance of 'std::invalid_argument'
what(): /usr/local/include/AnasaziBlockKrylovSchurSolMgr.hpp:266:
Throw number = 1
Throw test that evaluated to true: !_problem->isProblemSet()
Problem not set.
coud you give a piece of advice?
Sunghwan Choi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://software.sandia.gov/pipermail/trilinos-users/attachments/20130107/dca58656/attachment.html
More information about the Trilinos-Users
mailing list