Sender: root AT delorie DOT com
Message-ID: <388619D7.62566AFC@inti.gov.ar>
Date: Wed, 19 Jan 2000 17:08:55 -0300
From: salvador <salvador AT inti DOT gov DOT ar>
Organization: INTI
X-Mailer: Mozilla 4.6 [en] (X11; I; Linux 2.0.38 i686)
X-Accept-Language: es-AR, en, es
MIME-Version: 1.0
To: djgpp AT delorie DOT com
CC: Dieter Buerssner <buers AT gmx DOT de>
Subject: Re: gcc optimization (Was: Executable size: limit to acceptability?)
References: <Pine DOT SUN DOT 3 DOT 91 DOT 1000119113049 DOT 9609J-100000 AT is>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp AT delorie DOT com

Eli Zaretskii wrote:

> On 18 Jan 2000, Dieter Buerssner wrote:
>
> > My CPU is AMD K6-2 266.
>
> I don't know anything about K6.  AFAIK, GCC's code is optimized
> towards Intel's recommendations; I don't know how well these fit K6.
>
> > gcc 2.9.2: flags -fomit-frame-pointer -ffast-math + indicated flags
>
> You mean 2.95.2, right?
>
> >           -On -mcpu=k6  -On -march=k6
> > -O   86383       92070          92070
> > -O2  85852       86966          87009
> > -O3  81476       89791          89814
> > -O6  81421       89833          89818
> >
> > In all three cases -O produces the fastest code.
>
> The differences are small enough to be explained by alignment.

I agree with Eli. A very important point here is your CPU. AMD K6 CPUs have a
very huge problem with code aligned to 0x*C addresses. If the entry point of a
loop is there you are lost. As newer verions of gcc usually have better aligment
chances you hit a 0x*C (4 bytes aligned addres) increases.
Take a look to my http://www.geocities.com/SiliconValley/Vista/6552/compila.html
page, it have some interesting information about gcc optimizations.
Honza was doing a very good job adding optimizations for K6, but they are only
in the experimental branch of gcc.
I have a copy of gcc 2.96, it changes some interesting things in the
optimization, lamentably I didn't have the time to merge Honza patches to it and
see the performance increase but according to his reports K6 gets a very good
boost making gcc 2.96 the best compiler for K6 ;-)
You can play with the new switches found in gcc 2.95 to fine tune the aligment
(you can individually control loops, jumps, etc.).

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