RE: Pentium 4 Optimization Complimenting icc 5.0

["Walton, Shane" <swalton at dsrnet dot com>]

Thank you kindly.  I didn't mean to belittle gcc, I actually prefer it.
I am being forced to use the latest and (greatest?) of hardware, which
brought me to these questions.  I appreciate all the help.  Keep up the
good work all!  I look forward to gcc 3.1.

Shane M. Walton, Software Engineer
Digital System Resources, Inc.
swalton@dsrnet.com
703.234.1674


-----Original Message-----
From: Joe Buck [mailto:jbuck@synopsys.COM]
Sent: Wednesday, February 06, 2002 15:36
To: Walton, Shane
Cc: gcc@gnu.org
Subject: Re: Pentium 4 Optimization Complimenting icc 5.0



> I have been benchmarking Dual 1GHz Pentium 3 vs. Dual 1.7GHz Pentium 4
> Xeon systems with
> known UNIX/LINUX benchmark suites such as "ubench" and
"unixbenchmark".
> My findings are
> almost unbelievable.  The P3 system outperforms the P4 system by 20%.

You are using versions of gcc that were released before the P4 even
existed.  The P4 has a radically different pipeline structure than the
P3,
so code optimized for one doesn't work well with the other.
You can't expect the gcc team to predict the future.

Intel has been attacked for this, by the way: the P4 often is slower
than
the P3 for code tuned for the P3.

gcc 3.1 will contain decent P4 optimizations; you can test them out
by downloading a snapshot.  Release is scheduled for April, but that's
not a promise.