Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!ucsd!swrinde!gem.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!uunet!uvm-gen!wirthlin
From: wirthlin@uvm-gen.UUCP (Ralph Wirthlin,Gamma IV,31415926535,27182818)
Newsgroups: comp.binaries.ibm.pc.d
Subject: Re: /286 processor speeds:question
Message-ID: <1336@uvm-gen.UUCP>
Date: 15 Nov 89 14:34:28 GMT
References: <5392@cps3xx.UUCP>
Sender: nobody@uvm-gen.UUCP
Organization: EMBA Computer Facility, Univ. of Vermont, Burlington.
Lines: 41

From article <5392@cps3xx.UUCP>, by usenet@cps3xx.UUCP (Usenet file owner):
> In article <3492@vax1.tcd.ie> dbearly@vax1.tcd.ie (BIG BAD DOM) writes:
>>Could someone please tell me does a 286 @ 8Mhz run any faster then
>>an 808/86 @ 8MHz ?
> 
> Of course it does; at least twice as fast.  This is because the '286
> accesses memory 16 bits at a time (fetch), as opposed to 8 bits at a
> time (fetch fetch).  A '386 (not an sx) will access a full 32 bits at a
> time (FETCH!), and so if it runs at the same clock rate it will be
> roughly twice as fast again.  Of course, if you take advantage of the
> enhanced instruction set, you can get even greater speedup.

	An 8086 will also access memory 16 bits at a time on a word boundary
(fetch).  As someone pointed out, however, the 286 has reduced the
number of cycles required for a few of the, previously, expensive
instructions such as div or mul.  I also believe (don't have the
tech manual handy) that the pre-fetch queue on the 286 is larger than
the 8086, thus increasing the throughput.  
	The relative speed of the 286 over the 86 is obviously complicated by 
the type of program being run.  Is the program cpu or memory-bound?
I believe (again, no manual) that several of the string operations on
the 286 use the DMA controller.  If so, then a program that makes extensive
use of these instructions will run a good deal faster (about 4 times ?????).
	Conclusion?  Well, yes, an 8-Mhz 286 will generally run faster than an
8-Mhz 8086.  Relative speed?  That depends.  Try running norton on both.
( Someone correct me if I made any obvious mistakes :-)).

				Ralph

-------------------------------------------------------------------------------
Ralph L. Wirthlin          	    |  "Who knows where madness lies ..
Internet:wirthlin@emily.uvm.edu	    |   To surrender dreams, this may be 
BITNET: R_WIRTHLIN@UVMVAX	    |   madness .. too much sanity is madness,
				    |	but maddest of all is to see life as
USMAIL: 9B College Parkway	    |	it is and not as it should be"
	Colchester, VT 05446	    |              - Man of La Mancha	
-------------------------------------------------------------------------------