[petsc-users] unreliable AMG in PETSc

[Jed Brown]

Arne Morten Kvarving <arne.morten.kvarving at sintef.no> writes:

> On 23/10/14 15:53, Jed Brown wrote:
>
> hi jed,
>> Arne Morten Kvarving <arne.morten.kvarving at sintef.no> writes:
>> Do you mean if the matrix is _not_ marked SPD?  Interestingly, CG is
>
> no i mean to warn if the matrix is marked SPD, main KSP is CG and
> Richardson is used on the multigrid ksps. i guess it falls into user
> app space and is hard to do on a general level since, as you say, cg
> might survive slight asymmetries (or an user could be filtering out
> the naughty modes, like the hydrostatic).

There are many ways to get a preconditioner that is not SPD.  It's no
safer if the matrix is marked SPD.  And Richardson smoothing is not
inherently problematic; more like the relaxation paramater is easy to
mess up.
-------------- 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/20141023/4b3e48ce/attachment.pgp>