Message-ID: <B0000076659@stargate.astr.lu.lv>
From: "Andris Pavenis" <pavenis AT lanet DOT lv>
To: Nate Eldredge <nate AT cartsys DOT com>, djgpp AT delorie DOT com
Date: Wed, 24 Feb 1999 13:50:21 +0200
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: 586 to 386
In-reply-to: <36D32D50.585763E5@cartsys.com>
X-mailer: Pegasus Mail for Win32 (v3.01d)
Reply-To: djgpp AT delorie DOT com

On 23 Feb 99, at 14:36, Nate Eldredge wrote:

> Leif wrote:
> > 
> > Hi, im running DJGPP and RHide on a 586 processor.
> > I would like to compile my code to run on a 386.
> > Is there a way to tell the compiler not to use the any 586 optimizations
> > so I can get it to run on a 386?
> 
> Unless you specifically tell the compiler to do 586 etc optimizations,
> it won't, AFAIK.  If you're paranoid, you can use `-march=i386'.
> 
> See also FAQ chapter 11 if the 386 in question has no FPU.

Even more - seems that also binaries generated with -march=pentium still runs 
on 386:

	-  I have tested egcs-1.1.1 built for i586-pc-msdosdjgpp on 386 and it seems to
          work. Also binaries generated by it seems to be Ok. Only later I found that
           configure option --target=i586-pc-msdosdjgpp forces use of 
         -march=pentium -mcpu=pentium by default. Also I haven't heard about 
         related problems with egcs-1.1.1 binaries for DJGPP (they are built for 
         i586-pc-msdosdjgpp). 

        - I found that the in Slackware-3.5 and 3.6 distributions (Linux) the sitation is
          the same. -march=pentium -mcpu=pentium is used by default and seems
          that nobody complains. But I think there are poeple who installs Slackware
          3.5 or 3.6 on 386 or 486.

But anyway this thing can change in future.

Andris