Date: Sat, 21 Mar 1998 17:36:22 -0800 (PST)
Message-Id: <199803220136.RAA16746@adit.ap.net>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
To: James W Sager Iii <sager+@andrew.cmu.edu>, djgpp AT delorie DOT com
From: Nate Eldredge <eldredge AT ap DOT net>
Subject: Re: DJ: Compiler Optimizations
Precedence: bulk

At 09:50  3/21/1998 -0500, James W Sager Iii wrote:
>>John M. Aldrich wrote:
>                                                                             
>>Overall, the best optimization level for non-PGCC is -O2.  -O3 tries to       
>>perform inlining, which sometimes speeds code up but also occasionally 
>>slows it down.  -m486, -ffast-math, and -fomit-frame-pointer are also         
>>good optimization flags.  The gcc docs list all of them.                      
> 
>Ok, well you've been an incredible help. My compiler is already well over 4x 
>its previous speed.  I can't find the -m486 flag.  What menu is it under
>in rhide?

The info node, which you can also find through RHIDE's help, is "GCC"
"Invoking" "Submodel" "i386". Note that, as far as I know, `-m486' is not
advantageous unless the code will be run on a bona fide 486. It adds
alignment which helps the 486, but merely bloats the code on a 386 or
Pentium. (It's claimed that the Cyrix 6x86 also benefits from `-m486', but
on mine, very minimal tests show that it doesn't.)

Nate Eldredge
eldredge AT ap DOT net



 in
>the www.delorie.com site... but seems that link is outdated.
>will try digging up the sources on the linux distribution
>cd though. would that version work?

Probably not. Check the archives, somebody announced a version of vi for
DJGPP some time ago.

Nate Eldredge
eldredge AT ap DOT net