Error: L5 : elliptic PDE (PETSc) solver diverged. Reason: -9

Question

Hi

Before, I asked a very similar question, but then the Reason was -8, while now it is -9. The solution for -8 was to add a ground to my bidomain simulations. Now, I have a ground and I use the configure option "download-hypre" for the PETSc configuration, as was the answer for another similar question on this forum.

What is the problem here and how can I solve it?

Thank you a lot in advance!

Lore

Answer 1

Hi!

Based on: https://petsc.org/release/docs/manualpages/KSP/KSPConvergedReason.html

you seem to have NaN or Inf in your linear system. Most likely this is due to a bad input mesh.

With the latest commit, I have added a index-to-string conversion for the Converged/Diverged reason index of PETSc. This should help identifying future issues.

Best, Aurel