[petsc-dev] PETSc multi-GPU assembly - current status

Jed Brown jedbrown at mcs.anl.gov
Fri Jun 7 07:17:46 CDT 2013

Matthew Knepley <knepley at gmail.com> writes:

>> > For GPU only preconditioners,  would focus on the Cusp AMG using
>> > Chebychev for
>> > the smoothers.
>> OK. Again we'd have to create our own PCShell for this when using
>> MatShell if I understand correctly?
> I don't think so since Cheby just uses a matrix action.

Cheby only needs matrix action, but I don't understand the proposal to
use "Cusp AMG".

With a MatShell, you probably want to implement MatGetDiagonal to
correct the scaling, otherwise you'll be asking Chebyshev to do too

More information about the petsc-dev mailing list