[petsc-dev] Jacobian lag persist with ROSW method

[Peter Brune]

You can control Jacobian lagging with -snes_lag_jacobian and
-snes_lag_jacobian_presists by setting the option
-ts_rosw_recompute_jacobian to true.  Otherwise it defaults to having it
recompute once per timestep by setting the SNES to do so, wiping out the
persistent lagging.  This will give you a rather odd pattern of Jacobian
assembly, and I don't recommend it at all.  Your Jacobian will often be
computed on the unconverged nonlinear solution, and may be garbage.  What
would be more reasonable is for us to add some option at the TS level to
lag the Jacobian by a number of steps.

- Peter


On Sat, Mar 8, 2014 at 6:43 PM, Mani Chandra <mc0710 at gmail.com> wrote:

> Hi,
>
> The option -snes_lag_jacobian_persists TRUE doesn't seem to prevent the
> jacobian calculation at the beginning of each time step in the ROSW method
> in the TS module. I'm hoping to avoid an expensive jacobian assembly at the
> beginning of each time step and whenever a time step is rejected in the
> ROSW method. The option however works with the THETA method.
>
> I'm using the dev version.
>
> Cheers,
> Mani
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20140308/e4753b2e/attachment.html>