[francois pacull] Re: fieldsplits

[Jed Brown]

On Thu, 10 Dec 2009 15:12:38 -0600, Barry Smith <bsmith at mcs.anl.gov> wrote:
> 
> On Dec 10, 2009, at 3:03 PM, Jed Brown wrote:
> 
> > Can KSPView before KSPSolve reasonably be made to work?
> 
>     In general KSPView() should work perfect after KSPSetUp() but  
> before KSPSolve() BUT with the heirarchical preconditioners (like even  
> block Jacobi) some inner KSPSetUps() may not happen before the  
> KSPSolve() hence the view may not work correctly.
> 
> There is a KSPSetUpOnBlocks() that is suppose to make sure the inner  
> KSPSetUp()s are called, but with many-level heirarchical  
> preconditioners it may not get all the inner setups done properly.
> 
>    I do not know if it is possible to get it all correct in all cases  
> of composing preconditioners. I think  it is difficult.

I think so too, he said it "crashed", and I could produce SIGSEGV by
putting DMMGView before DMMGSolve in multiphysics/mp.c
(MatSchurComplementGetKSP called with A=0).  I just pushed a patch to
prevent fieldsplit from attempting to view inner objects before they
exist (just say "not yet available").

Jed