Date: Sun, 30 Nov 1997 16:09:10 -0800 (PST)
From: "William A. Barath" <wi534 AT victoria DOT tc DOT ca>
To: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
cc: "Alan M. Doerhoefer" <aland AT seanet DOT com>, djgpp AT delorie DOT com
Subject: Re: binutils 2.8.1
In-Reply-To: <Pine.SUN.3.91.971130131854.5422M-100000@is>
Message-ID: <Pine.GSO.3.95.iB1.0.971130160519.17505E-100000@vtn1>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk

On Sun, 30 Nov 1997, Eli Zaretskii wrote:

|On Thu, 27 Nov 1997, Alan M. Doerhoefer wrote:
|
|> In both cases, I used the same makefile, with these compile/link
|> switches:	-v -s -m486 -O3 -ffast-math
|
|Why -O3?  Did you try -O2?  If so, what were the results?

True enough.  Various things I have tried (such as the MAME drivers, parts
of DOSEMU, parts of the linux kernel) have produced both smaller and
faster code with -m386 and -O2 rather than -m486 and -O3.  And other
mixes.  This may not be the case with a Pentium.  I am using a Cyrix
6x86/PR120+ and it seems to ignore a lot of the rules for making code fast
on a pentium.  :)

Wil Barath, aka WseM : "I feel as though I see my pen to write"
Author of VPM, EDITPLN, and other VGA Planets support programs

<h3><hr><a HREF="http://victoria.tc.ca/~wi534">Visit my Homepage <a/><hr></h3>