request for advice cleaning up SNES

Barry Smith bsmith at mcs.anl.gov
Mon Sep 3 21:12:41 CDT 2007


  It is tentatively ok from me to remove them, there was
a systematic plan for these in the past (also for KSP) but
if that ever comes back we can always add them back again.

   Barry


On Mon, 3 Sep 2007, Lisandro Dalcin wrote:

> On 9/3/07, Matthew Knepley <knepley at gmail.com> wrote:
> > On 9/3/07, Lisandro Dalcin <dalcinl at gmail.com> wrote:
> > > In de process of cleaning up SNES, I am not sure about the management
> > > of afine (rhs) vector.
> >
> > I think it should be destroyed.
> 
> I think the same. I think I'll do that.
> 
> >
> > > BTW, I believe in the past you had and next removed
> > > KSPSetRhs()/KSPSetSolution() . Should this approach be also followed
> > > in SNES?
> >
> > Not sure. Why are these there?
> 
> I do not know. If you grep in petsc-dev,
> SNESSetRhs()/SNESSetSolution() are not used in any place other than
> SNES interface.
> 
> Do you think we can get rid of these two funtions and follow the
> behavior of KSP?
> 
> 
> 
> 
> 




More information about the petsc-dev mailing list