yy.wayne:
I'm using mumps and superludist through Trilinos' Amesos wrapper directly.
The code is simple, just step4 with TrilinosWrappers:SparseMatrix and
LinearAlgebra::distributed:Vector.
Here's the result from my virtual machine for a 98k DoFs problem.
Snipaste_2023-02-23_18-31-08.png
I've test a 300k+ DoFs problem with Amesos_mumps on server as well.
Multiple processes is not faster than one process.
It's not clear to me what you are doing. step-4 is not parallel, and unless
you extended it like we show in step-40, for example, all that happens when
you run the program with
mpirun -np 4 ./step-4
is that the same program is run four times on four different processes, but
all these four copies do is run the exact same instructions. This might also
explain why you see the output printed multiple times. If that is so, you
shouldn't be surprised that that takes exactly as long as running one instance.
If you want to test these kinds of solvers, start with step-40.
Best
W.
--
------------------------------------------------------------------------
Wolfgang Bangerth email: bange...@colostate.edu
www: http://www.math.colostate.edu/~bangerth/
--
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.
To view this discussion on the web visit
https://groups.google.com/d/msgid/dealii/00581715-f96e-0f90-e211-88899b058c73%40colostate.edu.
