On Tue, Sep 15, 2015 at 3:55 AM, Stephan Herb < inf74...@stud.uni-stuttgart.de> wrote:
> Thanks Julian for addressing this problem; although I changed it to the > correct version nothing changed (well, now only the displacements for > the local nodes gets printed instead of all nodes^^). > I came up with another issue: If you run the example code 4 of "Systems > of Equations" (Linear Elastic Cantilever) with the command "mpirun -np 2 > ./example-opt" one gets "-nan" as solutions. When you export it for > example as vtu files theone part cannot be loaded in a visualization > program like paraview and the other part can be loaded but only contains > zeros... >From your config.log file, I confirmed that you aren't using PETSc, and the sparse Eigen solvers are being used instead. Since the Eigen sparse solvers are serial, you can't use them with mpirun. I confirmed the result you described by running: mpirun -np 2 ./example-opt --use-eigen That is, half of the solution consists of NaNs while the other half is all zeros when you do this: https://drive.google.com/file/d/0B9BK7pg8se_ibkdpckxpVFI1LWM/view?usp=sharing So, your fix is to either run serially or install PETSc and make sure that libmesh configures and builds against it. We should probably try to prevent this type of configuration from getting run by throwing an error if the Eigen sparse solvers are invoked when n_processors() > 1. @roystgnr any preferences for how to do this? -- John ------------------------------------------------------------------------------ _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
