Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!princeton!caip!topaz!ll-xn!nike!ucbcad!ucbvax!PAVEPAWS.BERKELEY.EDU!dillon
From: dillon@PAVEPAWS.BERKELEY.EDU (Matt Dillon)
Newsgroups: net.micro.amiga
Subject: Re: 68010 & Math intensive stuff
Message-ID: <8606240057.AA09087@pavepaws>
Date: Mon, 23-Jun-86 20:57:22 EDT
Article-I.D.: pavepaws.8606240057.AA09087
Posted: Mon Jun 23 20:57:22 1986
Date-Received: Wed, 25-Jun-86 06:30:21 EDT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: University of California at Berkeley
Lines: 29


	Extracting entries from Motorola's programmer's ref. man for the
68000: (pg 189) note that this doesn't include EA calculation time, but
that's the same anyway between processors (or is trivial comparativly)


			68000		68010		speed increase
	
	MULS		70		42		1.6  x
	MULU		70		40		1.75 x
	DIVS		158		122		1.29 x
	DIVU		140		108		1.29 x


	The multiply and divide instructions have obviously been cleaned
up for the 68010 (most other instructions are the same, or within a clock,
however: mul/div enhancements are definately the cream of the crop in terms
of changes)

	For intensive math, 20% is reasonable considering that other 
instructions with less drastic changes between processors will also be
executed.

	However, I would test it out (with few exceptions, Never believe
your dealer)


				-Matt