Sender:	root AT delorie DOT com
Message-ID:	<378B8B2B.605BB1B2@inti.gov.ar>
Date:	Tue, 13 Jul 1999 15:53:31 -0300
From:	salvador <salvador AT inti DOT gov DOT ar>
Organization:	INTI
X-Mailer:	Mozilla 4.6 [en] (X11; I; Linux 2.0.36 i686)
X-Accept-Language:	es-AR, en, es
MIME-Version:	1.0
To:	djgpp AT delorie DOT com
Subject:	Re: gcc 2.96 benchmarks
References:	<378B57FE DOT 70A60D5C AT inti DOT gov DOT ar> <378B6AE6 DOT B7E12F66 AT eik DOT bme DOT hu>
Reply-To:	djgpp AT delorie DOT com

"Dr. András Sólyom" wrote:

> salvador wrote:
>
> > Hi All!
> >
> >   I tested the new experimental IA32 (it means x86) branch of gcc 2.96 with
> > the BYTE benchmarks.
> >   Here are some details:
> >
> > 1) MSVC stills the better, but not for much.
> > 2) 2.96 have the same speed for FPU than MSVC for Pentium MMX processors
> > (that's in average).
>
> Just my experiments with MSVC 5.0: without optimalization my code runs correctly
> (give same results as MATLAB), with optimization ON it gave incorrect results...
>
> Some months ago I read in DDJ (Dr. Dobbs Journal) an interview with the man who
> created or helped to create the first math co-processor. He said the the
> floating point optimizations in MSC are inferior and although they speeds things
> up, they lead to incorrect results. It seems to me he was right...

Is not the first time somebody points it, and coming from you I really think is
true. But even with it last snap shots are getting very close.
One comment I got is that MSVC is very far from compliant with the IEEE floating
point standards.
Jan did some interesting patchs to optimize even better and according to tests in
your machine with these patches gcc generates better code for K6 CPUs than MSVC and
the values I got with gcc 2.96 for Pentium MMX are very good in FPU. That's good:
better and with same speed ;-)

SET

--
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: set-soft AT usa DOT net set AT computer DOT org
                    set AT ieee DOT org set-soft AT bigfoot DOT com
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

- Raw text -