[Numpy-discussion] GMRES internal variables

David Huard david.huard@gmail....
Tue Feb 17 09:10:32 CST 2009


Nathan,

First of all, thanks to all your work on the sparse linear algebra package,
I am starting to use it and it's much appreciated.

Just a thought: wouldn't it be more natural to write gmres as a class rather
than a function ? That way, accessing the internal work arrays for reuse
would be much easier. These are useful when using gmres in a inexact newton
loop.

Cheers,

David




On Tue, Feb 17, 2009 at 2:15 AM, Nathan Bell <wnbell@gmail.com> wrote:

> On Mon, Feb 16, 2009 at 10:49 PM,  <jordan@math.ucsb.edu> wrote:
> >
> > I'm trying to run multiple instances of GMRES at the same time (one
> inside
> > another actually, used inside of the preconditioner routine) but am
> > running in to some problems. My guess is there is a single set of
> internal
> > variables associated with GMRES and when I initiate a new GMRES inside a
> > currently running GMRES, the old GMRES data gets the boot.
> >
> > Is this correct, and if so any suggestions on how to fix this?
> >
>
> This recently came up on SciPy-User:
> http://thread.gmane.org/gmane.comp.python.scientific.user/19197/focus=19206
>
> One solution:  PyAMG's GMRES implementation (pyamg.krylov.gmres)
> should be a drop-in replacement for SciPy's gmres().  Unlike the CG
> code mentioned above, you can't use gmres.py without some other
> compiled components (i.e. you'll need the whole pyamg package).
>
> We should have this resolved in the next scipy release (either 0.7.x or
> 0.8).
>
> --
> Nathan Bell wnbell@gmail.com
> http://graphics.cs.uiuc.edu/~wnbell/<http://graphics.cs.uiuc.edu/%7Ewnbell/>
> _______________________________________________
> Numpy-discussion mailing list
> Numpy-discussion@scipy.org
> http://projects.scipy.org/mailman/listinfo/numpy-discussion
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://projects.scipy.org/pipermail/numpy-discussion/attachments/20090217/4851b5bf/attachment-0001.html 


More information about the Numpy-discussion mailing list