[petsc-users] pcfieldsplit for a composite dm with multiple subfields

[Jed Brown]

Gideon Simpson <gideon.simpson at gmail.com> writes:

> The coming back to the original problem, I’m still unsure of how to diagnose the problem.  
>
> I am not sure if the listserv supports image attachments,

Yes, if small.

> but I dumped the snes_check_jacobian_view to disk and inspected the
> matrices with the spy command. I find that no entry in the saved
> matrix has absolute value larger than 10^{-2}.  There are some that
> have absolute value larger than 10^{-3}, but the actual values of the
> entries at those indices are on the order of 100 or 1000, meaning that
> they are small in a relative sense.

It's Frobenius norm and 0.000151055/6.08281e-10 = 248330.


You can try computing in quad precision, but it's possible that your
Jacobian is correct and your problem is just very nasty.  People usually
use continuations of various sorts for solving those.  For example, SNES
ex19 at high Grashof number will not converge even with grid sequencing,
but TS ex26 can solve the same problems handily.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 818 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20150909/58b178b8/attachment.pgp>