[petsc-users] Reuse MUMPS factorization
Matthew Knepley
knepley at gmail.com
Sat Mar 12 12:50:09 CST 2022
On Sat, Mar 12, 2022 at 1:01 PM Bhargav Subramanya <
bhargav.subramanya at kaust.edu.sa> wrote:
> Dear All,
>
> I have the following two queries:
>
> 1. I am running simulations using MUMPS through a job submission system.
> Since the job run time is limited, I need to restart the simulations
> periodically. Since I am restarting the simulation with exactly the same
> parameters, for instance, number of nodes, number of tasks per node, etc.,
> is it possible to dump the matrix factors for the first time and reuse it
> for the remaining restarts?
>
We do not have access to the MUMPS internals, so this is a MUMPS mailing
list question I think.
> 2. In the Ax=b solve, for now, the Matrix A doesn't change. However, if I
> implement a slightly different time integration technique, matrix A changes
> with each time step (sparsity pattern is still the same). Is it possible to
> incorporate a modifying matrix A in the mumps solver?
>
You can preserve the symbolic factorization, and just carry out the
numerical factorization. I believe we do this automatically if you just
update the values.
Thanks,
Matt
> Thanks,
> Bhargav
>
> ------------------------------
> This message and its contents, including attachments are intended solely
> for the original recipient. If you are not the intended recipient or have
> received this message in error, please notify me immediately and delete
> this message from your computer system. Any unauthorized use or
> distribution is prohibited. Please consider the environment before printing
> this email.
--
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
https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20220312/efb95ff9/attachment.html>
More information about the petsc-users
mailing list