[petsc-users] Tensor product as matrix free method

Mathis Friesdorf mathisfriesdorf at gmail.com
Tue Apr 1 09:47:23 CDT 2014

Thanks Jed, this is great. I was trying to implement what was proposed in
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.

Again, I really appreciate the help. Thanks, Mathis

On Tue, Apr 1, 2014 at 4:39 PM, Jed Brown <jed at jedbrown.org> wrote:

> Mathis Friesdorf <mathisfriesdorf at gmail.com> writes:
> > Thanks Jed! The libaries you are pointing to look very interesting
> indeed.
> > For the particular implementation I have in mind, I was hoping to get
> away
> > with something easier, as my time to work on this is a bit limited. All I
> > really need is a way to construct a block-diagonal matrix where each
> block
> > is given by the same matrix and all blocks are on the diagonal. Am I
> wrong
> > to assume that this should be possible to implement with TAIJ?
> That is exactly TAIJ.  How large are the diagonal blocks?  The
> implementation is intended to be fast only for small sizes, though I
> could make it good for large sizes as well.  In the special case of
>   (I \otimes S) X
> where I is the sparse identity and S is the dense diagonal block, this
> involves interpreting X as a matrix rather than a vector and using dgemm.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20140401/c7095bd9/attachment.html>

More information about the petsc-users mailing list