If the message does not display properly, pleaseClick HereHi,How is everything going? Are you ready to join us next month in Cape Town? CSP Focus South Africa 2014 must be a
Hello,
when I try to KSPSolve my matrix petsc prints out that a diagonal entry is
missing:
KSPSetOperators(solver, _matCLU.matrix, _matCLU.matrix );
KSPSolve(solver, vin.vector, vp.vector);
And petsc is perfectly right about that specific entry being zero. So this may
be more of a
PETSc has this “extra” requirement that you provide values for the entire
diagonal of the matrix, even those locations with zero. So make sure that you
use MatSetValues() to put a zero on each diagonal entry that has zero.
Also from below you are using ICC that is incomplete cholesky
Miguel Angel Salazar de Troya salazardetr...@gmail.com writes:
Hi all
I'm solving a discretized PDE with TS. Inside my MonitorFunction(), hooked
up with TSMonitorSet(), I process the solution at each time step to write
another PETSc Vec. Then, I write this Vec to a matlab binary file. The
Hi Matt,
Bumping this thread. I have a similar question. The lines you wrote for
testnum 39 have also worked perfectly for me. However, when I change a
couple parameters, namely not interpolating the mesh and using
-petscspace_order 1:
./ex12 -run_type full -refinement_limit 0.015625