From: "Duncan Coutts" <Duncan DOT Coutts AT dial DOT pipex DOT com>
Newsgroups: comp.os.msdos.djgpp
Subject: Assembly in DJGPP (MMX etc)
Date: Sun, 5 Sep 1999 15:12:35 +0100
Organization: UUNET WorldCom server (post doesn't reflect views of UUNET WorldCom
Lines: 14
Message-ID: <7qtv03$ck$1@lure.pipex.net>
NNTP-Posting-Host: userk983.uk.uudial.com
X-Trace: lure.pipex.net 936542019 404 193.149.73.51 (5 Sep 1999 14:33:39 GMT)
X-Complaints-To: abuse AT uk DOT uu DOT net
NNTP-Posting-Date: 5 Sep 1999 14:33:39 GMT
X-Newsreader: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

I know the standard gcc inline assembler uses the AT&T syntax rather than
the more common intel notation. I suspect (but an no sure) that the inline
assembler is not very recent and does not support new instructions eg MMX,
P6 or 3DNOW (I noticed that it doesn't even support segment override
prefixes! - see farptr.h).
My question therefore is what is the best way to include assembly (including
new instruction sets - MMX in particular) into my programs. Also, I don't
suppose there is any chance of inlining instead of linking?
My aim is to optimise the rendering and geometric transformation routines in
my 3D engine using MMX instructions.

Thanks in advance for any suggestions.