Dear Bruno,
If I solve a problem with one million DOFs on a cluster of 64 processors,
every processor will be dealing with around 15000 DOFs.
So you mean even in that case direct solver would be slow? I thought direct
solver would solve the problem only on those DOFs that it owns.
Best,
Hamed
On Friday, October 28, 2016 at 7:59:06 AM UTC-5, Bruno Turcksin wrote:
>
> Hamed,
>
> 2016-10-27 19:03 GMT-04:00 Hamed Babaei <hamedb...@gmail.com <javascript:>>:
>
> > I've been using the default one, Amesos_KLU, and I checked
> > Amesos_Superludist but trilinos doesn't recognize it.
> > when we installed dealii and its dependencies on the cluster, I think we
> > didn't uncomment the "#PACKAGES="${PACKAGES} once:superlu_dist" in
> candi.cfg
> > so it seems we need to install them again.
> Yes, Superlu_dist is a code independent of Trilinos. You will need to
> install it and reinstall Trilinos
>
> > I want to solve my problem with more than one million DOFs, So you mean
> > Amesos_Superludist will work for that many DOFs when running on cluster
> ?
> It will work but it will take a while. With that many DOFs, you should
> use an iterative solver. Superlu_dist is a lot faster than KLU and it
> will be good for the current size of your problem but when you have
> more than one million of DOFs it will be slow.
>
> Best,
>
> Bruno
>
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
