This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: Benchmarks

From: Scott Robert Ladd <coyote at coyotegulch dot com>
To: Robert Myers <rmyers1400 at attbi dot com>
Cc: gcc at gcc dot gnu dot org
Date: Wed, 18 Jun 2003 12:08:35 -0400
Subject: Re: Benchmarks
References: <3EF0853E.4070109@coyotegulch.com> <3EF08A91.9010504@attbi.com>

Robert Myers wrote:

I know it's a tease to say it (no time to get a web-site up), but nbench has yielded some interesting surprises about how gcc and icc do on P3 and P4 architectures in the various incarnations of both compilers.

I will look into it.

Relying on the closed source SPEC benchmarks is, well, <expletive deleted>. The open source community needs its own open source benchmarks.

Agreed.

I reject the SPEC benchmarks because of their closed-source nature. In the case of benchmarks (or science, for that matter), closed-source is unacceptable because there is no indepedent verification of methods.

If I can't publish the benchmark source code, I don't use the benchmark. While this may mean that I get piles of e-mail from people who perform their own tests, it also helps verify my own assumptions and methods.

--
Scott Robert Ladd
Coyote Gulch Productions (http://www.coyotegulch.com)

Follow-Ups:
- Re: Benchmarks
  - From: Joe Buck

References:
- Benchmarks
  - From: Scott Robert Ladd
- Re: Benchmarks
  - From: Robert Myers

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]