Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!wuarchive!uunet!proto!joe
From: joe@proto.com (Joe Huffman)
Newsgroups: comp.lang.c
Subject: Re: Bug in TC2.0 Optimization
Message-ID: <1991May05.233526.10084@proto.com>
Date: 5 May 91 23:35:26 GMT
References: <1991Apr22.130418.8079@uni-paderborn.de> <1991Apr28.220249.1@cc.helsinki.fi> <1991May03.104059.13090@cs.ruu.nl>
Organization: Prototronics @ Sandpoint, Idaho
Lines: 11

nevries@cs.ruu.nl (Nico de Vries) writes:

>On 8086 and 80286 processors xor dx,dx is in most cases faster than
>mov dx, 0h. On 80386 and i486 there is no reel reason to do this any more
>by the way.

Code size is smaller with either xor  dx,dx  or sub  dx,dx than with
mov dx,0.  Especially when in 386 native mode -- mov edx,0 uses 4 bytes for
the constant 0.
-- 
joe@proto.com