[petsc-dev] Mark, what's the rational for keeping the Chebyshev tuning inside GAMG?

[Tobin Isaac]

On Wed, Feb 25, 2015 at 12:13:15PM -0500, Mark Adams wrote:
> >
> >
> >
> > MatProjectDirichlet() would be useful here, too: it would allow the
> > user to inject the right modifications to the near nullspace.
> >
> 
> I'm not sure this is a good abstraction. These are coordinate
> transformations.  Here you just want to know if the value should be zero
> but in general you want a real transformation.
> 
> If the user rotates the local coordinate system to match the BC and zeros
> out row(s) then there will be a dummy equation and you should, in theory,
> rotate the null space, and the dumb MatProjectDirichlet do its job.  If the
> user _then_ rotates the operator back (what you initially suggested I
> think) then I don't think you need to do anything, now that I think about
> it.

I see your point.  Any users with local coordinate rotations for
Dirichlet conditions can probably provide their own near nullspace.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 490 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20150225/4142c9cb/attachment.sig>