> John Hearns wrote:
> > A commercial compiler - Intel / Pathscale /Portland will always give
> > you better performance from an application.
> > The guys who work on those compilers care about optimization.
> >
> >
> While I agree with this generality, generally. It is not always true
> Specific software may actually run faster with GCC than with a
> "optimized" compiler. We saw several instances of this on the Apple
> cluster with GCC 4 vs. Absoft.

Of course, Mike is right.  Having worked on the PathScale compiler, we 
occasionally had customers report that this code or that code was faster when 
compiled with GCC.  We would then work to catch-up or beat GCC.

And since this thread has been about OpenFOAM and has the word "newbie" in the 
subject...  My team at QLogic has experience with building OpenFOAM and both we 
and customers consider building it a "challenge."  It is oriented towards using 
GCC (sometimes particular versions of GCC), and if you can get it to build with 
that, I would not search hard for alternative compilers for that code.

-Tom

> 
> Also, a particular code may run much better with one "optimized"
> compiler than another. I have some that require PGI to run at all. Some
> that run faster with PGI. And, others run better with Intel tham PGI.
> 
> --
> Mike Davis                    Technical Director
> (804) 828-3885                        Center for High Performance Computing
> jmdav...@vcu.edu              Virginia Commonwealth University
> 
> "Never tell people how to do things. Tell them what to do and they will
> surprise you with their ingenuity."  George S. Patton
> 
> _______________________________________________
> Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin
> Computing
> To change your subscription (digest mode or unsubscribe) visit
> http://www.beowulf.org/mailman/listinfo/beowulf

_______________________________________________
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Reply via email to