Newsgroups: comp.parallel.mpi
From: Lars Rzymianowicz <lr@mufasa.informatik.uni-mannheim.de>
Subject: Re: NESL
Organization: Dept. of Computer Science, University of Mannheim, Germany
Date: Fri, 05 Jun 1998 08:52:38 +0200
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <357795B6.F36846BD@mufasa.informatik.uni-mannheim.de>

Guy Blelloch wrote:
> I rewrote your NESL code, admitatively with a bit more experience than
> you.
> My code is:
>   3) BETWEEN A FACTOR OF 15 and 40 times faster than yours when I run
>      both versions on the same machine and compiler
>   4) a factor of 80 times faster when I turn full checking off
>   5) runs on much larger problem sizes

Well done, Guy. What do we learn? "Better optimize your programmer
than your compiler" ;-) A factor 80 is really amazing...

Lars
-- 
email: lr@mufasa.informatik.uni-mannheim.de
Homepage: http://mufasa.informatik.uni-mannheim.de/lsra/persons/lars/