[petsc-users] TS and snes_grid_sequence

[Emil Constantinescu]

Francesco,

It could be a problem of ODE stability, error control, nonlinear solver 
(cold start), etc.

In addition to what's been recommended, I would suggest trying a 
different integrator with better properties: try

-ts_type arkimex -ts_arkimex_type 2e -ts_arkimex_fully_implicit

In case it's a cold-start problem, arkimex methods can provide warmer 
guesses: to do that try adding

-ts_arkimex_initial_guess_extrapolate

To take off the error controller, use -ts_adapt_type none

This should take care of the common things.

Emil

On 2/10/16 6:33 PM, Francesco Magaletti wrote:
> First of all thank you everybody for the fast replies.
>
> Barry,
> I think the solution 1 is a good suggestion but I confess I’m a bit scared of going deep into the modification of the ODE integrator.
> I’m now using TSCN, the classical Crank-Nicolson 2nd order full implicit time integration. Maybe I can try to implement your solution
> together with BEULER, as suggested by Hong.
> By the way the solution changes rapidly, but not so fast to justify that small timestep value. The space variation instead are really fast and this is the reason why I need so many grid points.
>
> Jed,
> I hope you could solve soon the problem with the FAS scheme.
>
> Thank you again
> Francesco
>
>