Segfault with WhitneyAVSolver and 2nd order elements

[wiesi]

I'm calculating magnetic fields with the WhitneyAVSolver using Elmer 6.2 (Rev: 5385) on Ubuntu 11.04 x86_64. The mesh is generated by Gmsh 2.5.0.
If I change the element order of a geometry-file (which works with 1st order elements) to 2nd order ones, Elmer crashes with a segfault, so there seems to be a bug hidden somewhere.

Here's the output snippet:

Code: Select all

MAIN: -------------------------------------
MAIN:  Steady state iteration:            1
MAIN: -------------------------------------
MAIN: 
DefUtils::DefaultDirichletBCs: Setting Dirichlet boundary conditions
DefUtils::DefaultDirichletBCs: Dirichlet boundary conditions set
CRS_IncompleteLU: ILU(0) (Real), Starting Factorization:
CRS_IncompleteLU: ILU(0) (Real), NOF nonzeros:    919722
CRS_IncompleteLU: ILU(0) (Real), filling (%) :       100
CRS_IncompleteLU: ILU(0) (Real), Factorization ready at (s):     0.04
      62 0.8251E-11
ComputeChange: NS (ITER=1) (NRM,RELC): ( 0.73219875      2.0000000     ) :: mgdynamics
Segmentation fault

I've included the necessary files in the attachment. bar_o1.msh is the Gmsh file with 1st and bar_o2.msh with 2nd order elements.

[Juha]

Hi,
this solver won't work with 2nd order elements, this has not been implemented.
i'll add some cheks to exit early, thanks for the report.
Regards, Juha