[petsc-users] solving multiple linear systems with same matrix (sequentially, not simultaneously)

[Daniel Goldberg]

Thanks Jed.

So the nullspace would be initialized with an array of 3 petsc vectors:
(u,v)= (0,1), (1,0), and (-y,x), correct?

And also to be sure -- this is usefuil only for multigrid preconditioners,
yes?

Thanks
Dan

On Mon, Feb 23, 2015 at 2:11 AM, Jed Brown <jed at jedbrown.org> wrote:

> Barry Smith <bsmith at mcs.anl.gov> writes:
> >> I will try gamg as I know it can reduce the number of CG iterations
> required. (I'm guessing you mean algebraic, not geometric?)
> >
> > By default GAMG is an algebraic multigrid preconditioner. Look at its
> documentation at http://www.mcs.anl.gov/petsc/petsc-master/docs/index.html
> it will be a bit better than the older documentation. The documentation for
> GAMG is still pretty thin so feel free to ask questions.
>
> You might want to call MatSetBlockSize and MatSetNearNullSpace to define
> both translation and rotation modes.  This is most relevant for large
> ice shelves.
>



-- 

Daniel Goldberg, PhD
Lecturer in Glaciology
School of Geosciences, University of Edinburgh
Geography Building, Drummond Street, Edinburgh EH8 9XP


em: D <dgoldber at mit.edu>an.Goldberg at ed.ac.uk
web: http://ocean.mit.edu/~dgoldberg
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20150223/fbf35cee/attachment.html>