lapackers

[Jakub Kurzak <kurzak@eecs.utk.edu>]

It really is a nice post.
A long time ago it crossed my mind to make the following experiment.
- Take a Playstation 3,
- Run code from Numerical Recipes on the PowerPC core,
- Run LAPACK on the PowerPC core,
- Run hand-optimized code on the SIMD cores.
I am quite sure I could show 4 orders of magnitude difference.
This is a transition from something running roughly one year
to something running roughly one hour.
I think it tells us something about the state of computer science today :)
Jakub

On Wed, Jul 15, 2009 at 9:28 AM, James Demmel <demmel@cs.berkeley.edu> wrote:

Julien Langou wrote:
> Dear lapackers,
> let me point to you the following post on our forum:
> http://icl.cs.utk.edu/lapack-forum/viewtopic.php?f=2&t=1561
> --julien.
> _______________________________________________
> Lapackers mailing list
> Lapackers@cs.utk.edu
> http://lists.eecs.utk.edu/mailman/listinfo/lapackers
>

This is a good example to mention in my short course next week!
Do we know how big the matrix was, or what platform it ran on?
Jim

_______________________________________________
Lapackers mailing list
Lapackers@cs.utk.edu
http://lists.eecs.utk.edu/mailman/listinfo/lapackers