dolfin team mailing list archive

["Garth N. Wells" <g.n.wells@xxxxxxxxxxxxxxx>]

Quoting "Robert C. Kirby" <kirby@xxxxxxxxxxxx>:

> I'm skeptical about this approach of requiring HYPRE --
> it tends to violate the spirit of PETSc.
> The point of PETSc is that everything can be run from the command line
> (or a shell script that has all the options).  Somebody might refuse to
> download HYPRE because they won't register (I can't imagine anybody
> with that attitude, but I've heard they exist) or because they have  
> something
> better or whatever.
> 
> Perhaps a nicer thing to do is have a script that checks for HYPRE  
> and runs DOLFIN
> accordingly, or else does some default thing.

I agree, and this is how it works. It will still compile and run if you don't
have HYPRE, but'll you'll be warned that that the solver will be slow.

Garth

> 
> On Oct 28, 2005, at 8:57 AM, Anders Logg wrote:
> 
> > On Fri, Oct 28, 2005 at 11:12:25AM +0200, dolfin@xxxxxxxxxx wrote:
> >
> >> Commit from garth (2005-10-28 11:12 CEST)
> >> -----------------
> >>
> >> Add HYPRE check in NavierStokesSolver.cpp. If HYPRE has not been
> >> linked with PETSc, a warning is printed and the default GMRES solver
> >> is used for the continuity equation.
> >>
> >
> > Very good!
> >
> >
> >> Should add check for all non-standard PETSc components.
> >>
> >
> > I agree.
> >
> > /Anders
> >
> > _______________________________________________
> > DOLFIN-dev mailing list
> > DOLFIN-dev@xxxxxxxxxx
> > http://www.fenics.org/cgi-bin/mailman/listinfo/dolfin-dev
> >
> 
> Rob Kirby
> 
> "Mathematical software should be mathematical."
> 
> 
> 
> 
> _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/cgi-bin/mailman/listinfo/dolfin-dev
>