[petsc-dev] PCASM: subdomains spanning multiple processes

[Jed Brown]

On Tue, 6 Apr 2010 16:12:31 -0500, Barry Smith <bsmith at mcs.anl.gov> wrote:
> > Note that the present code has no special cases to
> > distinguish (1) and (2) above.
> 
>     There is PCSetUp_BJacobi_Singleblock()   
> PCSetUp_BJacobi_Multiblock(), destroy single block and  multiblock()  
> and apply single block and multiblock for bjacobi most of this is to  
> handle the special case where one does not need to make a COPY of the  
> matrix (since [SB]AIJ matrices have the "diagonal" block already).

Ah, I was only referring to the ASM code, I haven't looked at BJacobi
carefully.  I think (hope) that there is a cleaner way to support this.
I see that PCModifySubMatrices is not called in bjacobi single-block
(this prevents a copy).

Jed