[petsc-dev] [petsc-users] Recursive fieldsplit

[Matthew Knepley]

On Mon, May 20, 2013 at 11:35 AM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> Matthew Knepley <knepley at gmail.com> writes:
>
> > I think we have to allow an argument to the strided operations, but the
> arg
> > should be a factor of the bs of the Vec I think.
>
> What if the Vec has bs=1?
>
> Maybe it should either be a factor or a multiple.  So you can't use
> stride=3 on a bs=5 Vec or vice versa, but you can use stride=2 on a
> stride=6 Vec and vice versa.
>

Actually I now think I was wrong. Let me try to go through the use case,
and you tell
me where I am wrong:

  1) We start with a system with block size B

  2) We split it into k parts, so that \sum_k b_k = B, and b_k is the
blocksize of each
      component. Isn't this how FS for structured grids works?

So the condition should be any bs <= Vec bs, and we also check that it
divides the size.

    Matt

-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20130520/76e4084c/attachment.html>