From: Hans-Bernhard Broeker <broeker AT physik DOT rwth-aachen DOT de>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Voodoo optimization?
Date: 6 Jul 2000 10:58:38 GMT
Organization: Aachen University of Technology (RWTH)
Lines: 18
Message-ID: <8k1oou$a9i$1@nets3.rz.RWTH-Aachen.DE>
References: <20000706015618 DOT 01353 DOT 00001528 AT ng-bj1 DOT aol DOT com>
NNTP-Posting-Host: acp3bf.physik.rwth-aachen.de
X-Trace: nets3.rz.RWTH-Aachen.DE 962881118 10546 137.226.32.75 (6 Jul 2000 10:58:38 GMT)
X-Complaints-To: abuse AT rwth-aachen DOT de
NNTP-Posting-Date: 6 Jul 2000 10:58:38 GMT
Originator: broeker@
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

S. T. L. <stl137 AT aol DOT com> wrote:
[...]
> the commandline GCC -s -O2 -march=i686 -fomit-frame-pointer
> -ffast-math -funroll-loops -funroll-all-loops -malign-double
> -fstrict-aliasing -o BLADE.EXE *.c

Be careful with -fstrict-aliasing. It may generate code that doesn't
work reliably, as it makes assumptions about the code's behaviour that
many existing sources may not respect.

And, if -funroll-all-loops really gains something, then you should
probably also be using -O3 (or -finline-functions). Both have a
similar behaviour: they generate larger code, by repeating code
sections instead of looping or calling subprocedures, respectively.

-- 
Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
Even if all the snow were burnt, ashes would remain.