Re: New procedure bicgstab

[David Bateman]

Jaroslav Hajek wrote:
> On Mon, Dec 22, 2008 at 2:02 PM, Søren Hauberg <address@hidden> wrote:
>   
> > man, 22 12 2008 kl. 11:05 +0100, skrev Jaroslav Hajek:
> >     
> > > On Mon, Dec 22, 2008 at 10:14 AM, R! <address@hidden> wrote:
> > >       
> > > > Hi,
> > > >
> > > > I understand that good preconditioning is very important for procedure like
> > > > bicgstab and cgs which I wrote, so maybe I can now try to write ILU
> > > > procedure for Octave if you find it useful.
> > > >
> > > > But I'm also interested in your opinion about bicgstab now.
> > > >
> > > >         
> > > Certainly it's useful. Not even in the sparse world is every matrix
> > > ill-conditioned.
> > > The Krylov solvers are IMHO an important missing piece in Octave. It's
> > > great someone took up the job. Even not-quite-optimized
> > > implementations are a good take-off.
> > > Having gmres would be great, too.
> > >       
> > Where should we put stuff like this? The 'optimization' package is
> > option.
> >
> >     
>
> Radek's previous cgs contribution went into script/sparse/ and I guess it fits.
>   

That is where matlab puts them...  I don't have much time to deal with 
reviewing and adding this code to Octave.. It should go in the core as 
it is a core matlab function...

D.



--
David Bateman                                address@hidden
35 rue Gambetta                              +33 1 46 04 02 18 (Home)
92100 Boulogne-Billancourt FRANCE            +33 6 72 01 06 33 (Mob)