dolfin team mailing list archive
-
dolfin team
-
Mailing list archive
-
Message #00130
Re: Re: Linear Algebra
-
To:
Discussion of DOLFIN development <dolfin-dev@xxxxxxxxxx>
-
From:
Anders Logg <logg@xxxxxxxxx>
-
Date:
Tue, 26 Oct 2004 10:54:03 -0500
-
Cc:
Discussion of DOLFIN development <dolfin-dev@xxxxxxxxxxxxxxxxxxxxxx>, claes@xxxxxxxxxxxxxxxx, "Robert C.Kirby" <kirby@xxxxxxxxxxxxxxx>, Erik Svensson <eriksv@xxxxxxxxxxxxxxxx>, Matthew Knepley <knepley@xxxxxxxxxxx>, Ridgway Scott <ridg@xxxxxxxxxxxxxxx>
-
In-reply-to:
<y8hekjlwp5x.fsf@harley.mcs.anl.gov>
-
Mail-followup-to:
Discussion of DOLFIN development <dolfin-dev@xxxxxxxxxx>, "Robert C.Kirby" <kirby@xxxxxxxxxxxxxxx>, Erik Svensson <eriksv@xxxxxxxxxxxxxxxx>, Discussion of DOLFIN development <dolfin-dev@xxxxxxxxxxxxxxxxxxxxxx>, Ridgway Scott <ridg@xxxxxxxxxxxxxxx>, claes@xxxxxxxxxxxxxxxx, Matthew Knepley <knepley@xxxxxxxxxxx>
-
Reply-to:
Discussion of DOLFIN development <dolfin-dev@xxxxxxxxxx>
-
User-agent:
Mutt/1.5.6+20040907i
On Tue, Oct 26, 2004 at 09:05:14AM -0500, Matthew Knepley wrote:
> > Both/and. First, there is the mathematical issue (this is
> > tantamount and has the biggest impact). But even if you have an
> > optimal preconditioner, a crappy implementation of the
> > matrix-vector product (e.g. (i,j) index storing by unsorted
> > lists), you will have a huge difference in performance among
> > implementations.
> I wasn't allowing for anything that crappy. Most common
> implementations are fine, but someone could still screw it up.
Thanks, Matt. ;-)
/Anders
References
