[petsc-dev] Performance of Petsc Cholesky vs LU

[Jed Brown]

On Wed, Nov 30, 2011 at 13:38, Barry Smith <bsmith at mcs.anl.gov> wrote:

> Jed if you know how to do efficient Cholesky with nd ordering using the
> sbaij data structure that would be a great addition to PETSc but I think it
> is likely impossible.


You can build the column-oriented indices or just make a copy. I don't know
what CHOLMOD and MUMPS do, but even a full copy followed by factorization
in the new natural ordering isn't going to be a big deal. We wouldn't need
iterative solvers at all if direct solvers had a fill factor of order one.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111130/d0164a09/attachment.html>