[petsc-users] Tensor product as matrix free method

[Mathis Friesdorf]

Alright, so your point is that while this construction is in principal
possible, it will not be efficient. Thanks for clearing this up!


On Tue, Apr 1, 2014 at 4:57 PM, Jed Brown <jed at jedbrown.org> wrote:

> Mathis Friesdorf <mathisfriesdorf at gmail.com> writes:
>
> > Thanks Jed, this is great. I was trying to implement what was proposed in
> >
> http://lists.mcs.anl.gov/pipermail/petsc-users/2011-September/009991.html,
> > by you as I just realised. For this the matrices would be of varying
> size,
> > so for three local systems, I would need:
> >
> > S \otimes Id \otimes Id ; Id \otimes S \otimes Id ; Id \otimes Id
> \otimes S.
> >
> > Realistic systems are of the order of at least 20 local systems, which
> > would mean that I need the matrices of the above structure for all 20
> > possible positions.
>
> This is something for which you should use CTF or TCE.  When S moves out
> of the last position in the tensor product, the operation becomes less
> local and various forms of transposition is needed.  This is not
> something worth recreating in PETSc unless there is a critical
> deficiency in the available libraries.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20140401/6d462543/attachment.html>