On Mon, Apr 6, 2015 at 11:10 AM, Roy Stogner <royst...@ices.utexas.edu>
wrote:
>
> On Mon, 6 Apr 2015, David Knezevic wrote:
>
> PETSc will automatically avoid recomputing preconditioners if you call
>> solve multiple times with the same matrix and different right-hand-sides,
>> but we don't currently get this benefit because in PetscLinearSolver we
>> re-initialize the solver's matrix on each call to solve.
>>
>
> Copying to Vikram so he can help with my fuzzy memory.
>
> Didn't we already add LinearSolver::reuse_preconditioner() for this?
> Or are there additional savings possible in the case where you know
> that not only is the old preconditioner good enough but the old matrix
> is unchanged?
>
Yes, we did add reuse_preconditioner, and that essentially achieves the
same thing. I use that option all the time, and it works well.
So you're right, there's not really any need to add this
new solve_with_same_matrix method. I just thought it might be a bit easier
for users to understand, but it's fine with me if we don't bother adding it.
David
------------------------------------------------------------------------------
BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
Develop your own process in accordance with the BPMN 2 standard
Learn Process modeling best practices with Bonita BPM through live exercises
http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF
_______________________________________________
Libmesh-devel mailing list
Libmesh-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-devel
