Date: Thu, 28 May 1998 01:57:06 +0200 (SAT) From: Robert Nordier <rnordier@nordier.com> To: eivind@yes.no (Eivind Eklund) Cc: hasty@rah.star-gate.com, toor@dyson.iquest.net, current@FreeBSD.ORG Subject: Re: Fix for undefined "__error" and discussion of shared object versioning Message-ID: <199805272357.BAA16858@ceia.nordier.com> In-Reply-To: <19980527231803.18654@follo.net> from Eivind Eklund at "May 27, 98 11:18:03 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Eivind Eklund wrote:
> On Wed, May 27, 1998 at 02:07:55PM -0700, Amancio Hasty wrote:
> > Next, is can you come up with a simple benchmark to test the floating
> > point optimization for Pentium classes or higher?
>
> Not easily, no. Robert Nordier had som FP code he was compiling with
> it. I have no floating point intensive code, and I run on an atypical
> platform (the Pentium Pro, which has a quite different profile than
> the Pentium. The P-II is supposedly like the Pentium).
>
> >From looking at the optimizations passes I can say that it should be
> at least as fast as gcc for 386, 486, and Pentium Pros, faster for
> Pentiums, and I don't know for P-IIs - I have never really
> experimented to find the performance profile for the P-IIs.
Just as an experiment, I compiled an old grep utility using both GCC
and TenDRA. The results were really disappointing. :-(
The TenDRA-compiled utility runs around twice as slowly:
gcc 5.74 real 5.69 user 0.04 sys
tcc 10.64 real 10.48 user 0.07 sys
>From a 233 MHz PII running -current searching on /usr/share/dict/web2.
The files are at http://www.icon.co.za/~nordier/re.tar.gz
--
Robert Nordier
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199805272357.BAA16858>