[petsc-users] Changing nonzero structure and Jacobian coloring

[Ellen M. Price]

Hi PETSc users!

I have a problem that I am integrating with TS, and I think an implicit
method would let me take larger timesteps. The Jacobian is difficult to
compute, but, more importantly, the nonzero structure is changing with
time, so even if I use coloring and finite differences to compute the
actual values, I will have to update the coloring every time the
Jacobian recomputes.

Has anyone done this before, and/or is there a correct way to tell TS to
re-compute the coloring of the matrix before SNES actually computes the
entries? Is this even a good idea, or is the coloring so expensive that
this is foolish (in general -- I know the answer depends on the problem,
but there may be a rule of thumb)?

Thanks,
Ellen Price