On Thu, Oct 21, 2010 at 23:56, Matthew Knepley <knepley at gmail.com> wrote:
> You SNES is not converging. You solve 163 systems, but do 4300 function
> evaluations.

That is because the Jacobian is assembled using coloring.


Sundials is probably doing fewer steps than BEuler for this problem,
and it's also lagging the preconditioner and just using matrix-free
differencing to define the Krylov operator.  These options are not
default with PETSc's TS implementations.

Jed

Reply via email to