[petsc-dev] [petsc-maint] SNESMF

[Jed Brown]

Barry Smith <bsmith at mcs.anl.gov> writes:
>    Hmm, so if this is a reasonable thing (Jed and Matt?) shouldn’t we
>    add support for it for TSPSEUDO? 

TSPSEUDO supports something similar, but the residual is always the
steady-state residual.  That is actually the same thing you get if you
try linearly-implicit Euler and do not extrapolate, because the time
derivative is 0 each time you evaluate a residual.

Kostas, do you intend to perform a full nonlinear solve for a finite
time step, or leave the pseudotransient continuation to deal with
globalization?

I think we need a test case in which this beats TSPSEUDO to justify
adding it as an option.

>    This is also much like a trust region in that you are adding some
>    scaling of the identity (or some other matrix) to the Jacobian,
>    maybe.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20140213/35378848/attachment.sig>