[petsc-users] Simple diffusion problem: TSARKIMEX vs TSROSW vs TSBEULER

[Christophe Ortiz]

Hi all,

I have implemented some simple diffusion problem in 1D using PETSc.
I use TSSetIFunction and TSSetIJacobian.

My initial solution is a simple gaussian with a maximum peak at 1e18.

- When I use TSROSW, everything's fine. Timestep is adaptive and the
solution is correct (a gaussian)


- When I use TSARKIMEX, the timestep is adaptive but final solution is
completely wrong. It gives some strange profiles on boundaries with very
high values. Any idea why ?


- When I use TSBEULER, timestep remains constant. It takes the initial
timestep value given with TSSetInitialTimeStep and stays like that for
ever. This makes simulations never finish. Actually, I observed this
behavior with TSBEULER with the different examples given in tutorials. Why
is the timestep not adaptive with TSBEULER ? Is there a way to have it
adaptive ?

Many thanks in advance !
Christophe
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20130926/2cf04675/attachment.html>