This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
RE: GCC vs Intel C++ benchmarks
- From: "Scott Robert Ladd" <scott at coyotegulch dot com>
- To: "Paul Rubin" <phr-2002 at nightsong dot com>, <bug-gcc at gnu dot org>
- Date: Wed, 18 Dec 2002 14:17:43 -0500
- Subject: RE: GCC vs Intel C++ benchmarks
> The most interesting comparison is the Monte Carlo integration in
> Scimark 2.0. The code from Intel C++ is almost 3x faster on the
> Pentium III and almost 5x faster on the Pentium 4. On most other
> benchmarks, the Intel compiler's code is a little bit faster, not
> enough to get too worked up about.
As a couple of people have pointed out to me in e-mail, the MonteCarlo test
shows the advantages of Intels interprocedural optimizations. Vassilli
Karpov told me the following:
> MonteCarlo.c gets a needed boost if we do(ICC does not need this
> due to IPO):
> ==
> 1c1
> < #include "Random.h"
> ---
> > #include "Random.c"
> ==
>
> Ofcourse runtest.sh should be modified to not pass Random.c to gcc.
By the first of the year, I'll have a couple new "real world" benchmarks
online which will tell a more accurate story.
As always: If anyone associated with gcc has suggestions for improving its
performance, I'm listening!
..Scott