On Tue, 31 Mar 1998, Project ELLIPTIC/Paul Walker wrote:

>   
>  These conversations bring up an interesting thought; we have many linear 
>  elliptic solvers (but the three ones we really think about are BAM PETSc 
>  and SOR). Wouldn't it be nice if we could do 
>   
>  maximal_solver = "BAM_precond petsc" 
>   
>  or 
>   
>  maximal_solver = "BAM sor_postcond"     (which is what Mark wants) 
>   
>  to use BAM as a pre-conditioner to PETSc or SOR as a post condition to 
>  BAM. 
>   
>  Anyone in cactus land want to think about this? Seems it would be a change 
>  to linear elliptic solve (basically a parser) and a change to each of the 
>  elliptic solvers (they have to realize they are called as a pre- or post- 
>  conditioner and therefore behave differently, and we have to define how). 
>   
>  Just randomly inventing features for other people to implement :-) 
>   
>          - Paul 
>   

Yes, it would be very useful to have a simple way to pre-condition and
post-condition, even though proper pre-conditioning for example should be
done with the same stencils that are later used in, say, petsc. We have
discussed it before, but now I think we have a good reason to do it. I'll
put it on my to-do list, but anybody is welcome to just do it. 

BAM_Elliptic has a parameter bam_fullV which has to be set to "no", other
wise BAM does not use the old field as initial guess. For some reason
maximal slicing works better with "yes", which is the default.

Bernd

Created for the WUGRAV CoCoBoard Page Projects Page.

Created by The CoCoBoard.