From:	Charles Sandmann <sandmann AT clio DOT rice DOT edu>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Generate real-mode 386 DOS executables?
Date:	Mon, 13 Jan 2003 10:54:38 CST
Organization:	Rice University, Houston TX
Lines:	12
Message-ID:	<3e22ef4e.sandmann@clio.rice.edu>
References:	<xruU9.23044$jM5 DOT 60769 AT newsfeeds DOT bigpond DOT com>
NNTP-Posting-Host:	clio.rice.edu
X-Trace:	joe.rice.edu 1042477143 8881 128.42.105.3 (13 Jan 2003 16:59:03 GMT)
X-Complaints-To:	abuse AT rice DOT edu
NNTP-Posting-Date:	13 Jan 2003 16:59:03 GMT
X-NewsEditor:	ED-1.5.9
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

> Hmm. Is the 32-bit real mode that Charles was talking about where the
> address space is 32-bit, or where the operand size is 32-bit, or both?

I referring to operand size.  For example, BCC 5.0 supports a /3 switch
which generates code which will only run on 386+ systems.  It uses size
overrides in real mode to support using 32-bit registers, operations and
moves.  This is faster than doing it in multiple 16-bit chunks, but
not as fast as being in 32-bit mode to begin win.

The original post indicated he was happy with 640K but wanted 32-bit
support.  It wasn't clear from that if he needed only 32-bit manipulations
or simpler addressing.  I think he just wanted faster code ...

- Raw text -