Path: utzoo!attcan!uunet!bu.edu!rpi!zaphod.mps.ohio-state.edu!sdd.hp.com!elroy.jpl.nasa.gov!ames!ptolemy!eos!shelby!portia.stanford.edu!dhinds
From: dhinds@portia.Stanford.EDU (David Hinds)
Newsgroups: comp.lang.pascal
Subject: Re: 286- and 386- native code
Message-ID: <1990Jul17.211928.22498@portia.Stanford.EDU>
Date: 17 Jul 90 21:19:28 GMT
References: <10330@chaph.usc.edu> <BIANCO.90Jun21151125@zephyr.cs.odu.edu> <22741@boulder.Colorado.EDU>
Organization: AIR, Stanford University
Lines: 26

In article <22741@boulder.Colorado.EDU> streich@boulder.Colorado.EDU (Mark Streich) writes:
>In article <BIANCO.90Jun21151125@zephyr.cs.odu.edu> bianco@cs.odu.edu (David J. Bianco) writes:
>>
>>	   When does Borland plan to give us a Turbo Pascal which
>>	   generates 386 code?
>>
>>Soon, I would suspect, if not already. Turbo C++ supports 8086, 80186, 
>>80286, and 80386 native modes.
>
>What type of code does Turbo C++ generate for the 386?  Does it use the
>extended registers or what? (sorry, I know this is a pascal group...)

    I should say that this is just an educated guess, because I don't
have TC++.  I have a strong suspicion that the 386 support is limited to
extra instructions (bit field manipulation, procedure ENTER/LEAVE, etc)
and makes no use of 32-bit registers or extended addressing modes.  If
TC++ does generate 32-bit code, I will be very surprised; that would be
a MAJOR feature.  TC++ is a real-mode compiler - I am sure it will not
generate native (protected) mode 80386 code.  It is entirely possible
to generate 32-bit code that runs in real mode - it just takes lots of
size override prefixes.  I am unaware of any DOS compiler that does it,
though.  It should be a big plus for math-intensive code, so it is a
shame that it isn't done.

 -David Hinds
  dhinds@popserver.stanford.edu