Hello, I am getting the runtime error below in a simulation within a deal.II-based CPFE framework.
*Error message:* *----------------------------------------------------* *Exception on processing: * *--------------------------------------------------------* *An error occurred in line <2204> of file <../source/lac/lapack_full_matrix.cc> in function* * void dealii::LAPACKFullMatrix<number>::compute_eigenvalues_symmetric(number, number, number, dealii::Vector<OtherNumber>&, dealii::FullMatrix<number>&) [with number = double]* *The violated condition was: * * info == 0* *Additional information: * * Lapack error in syevx: 1 eigenvectors failed to converge. (You may* * need to scale the abs_accuracy according to your matrix norm.)* *Stacktrace:* *-----------* *#0 /usr/local/lib/libdeal_II.so.9.5.0: dealii::LAPACKFullMatrix<double>::compute_eigenvalues_symmetric(double, double, double, dealii::Vector<double>&, dealii::FullMatrix<double>&)* *#1 ./main: crystalPlasticity<3>::reorient2(dealii::Vector<double>&, dealii::Vector<double>, dealii::FullMatrix<double>, dealii::FullMatrix<double>)* *#2 ./main: crystalPlasticity<3>::calculatePlasticity(unsigned int, unsigned int, unsigned int)* *#3 ./main: crystalPlasticity<3>::getElementalValues(dealii::FEValues<3, 3>&, unsigned int, unsigned int, dealii::FullMatrix<double>&, dealii::Vector<double>&)* *#4 ./main: MultiPhysicsBVP<3, 1>::assemble()* *#5 ./main: MultiPhysicsBVP<3, 1>::solveNonLinearSystem()* *#6 ./main: MultiPhysicsBVP<3, 1>::solve_cp()* *#7 ./main: crystalPlasticity<3>::run()* *#8 ./main: main* *--------------------------------------------------------* *Aborting!* *----------------------------------------------------* Some relevant info is I am using a docker container with deal.II version 9.5.0 If I run on deal.II 9.4.2 with lapack v 3.11.0 installed externally in a cluster, the code throws the error:* LAPACK error in syevx,* but does not crash. The function *compute_eigenvalues_symmetric* is called with the following arguments: C_old.compute_eigenvalues_symmetric(0.0,200000.0,1e-15,eigenvalues, eigenvectors); I have also tried the following values for abs_accuracy: 1) 10e-16 2) 2 x 2.22507e-308 = 4.45014e-308 (twice the underflow threshold) 3) 0 In addition, I tried increasing upper_bound ny 3 orders of magnitude: 200000000.0. But all of these simulations crashed at some point. Any ideas for how I can solve this problem? Thank you, David -- 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 dealii+unsubscr...@googlegroups.com. To view this discussion visit https://groups.google.com/d/msgid/dealii/0d504d69-ec7b-4b0d-bc59-ef678ffc744fn%40googlegroups.com.
