Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version nyu B notes v1.5 12/10/84; site acf4.UUCP
Path: utzoo!linus!philabs!cmcl2!acf4!tsc2597
From: tsc2597@acf4.UUCP (Sam Chin)
Newsgroups: net.arch
Subject: Differences between 8088 and 80286 ?
Message-ID: <290008@acf4.UUCP>
Date: Sat, 1-Jun-85 20:37:00 EDT
Article-I.D.: acf4.290008
Posted: Sat Jun  1 20:37:00 1985
Date-Received: Sun, 2-Jun-85 20:19:15 EDT
Organization: New York University
Lines: 22

<>

  Recently there has been some discussion on net.micro about a 8 Mhz 8088
being faster than a 6 Mhz 80286.  In particular, it was a comparison of
benchmarks written in diffrent dialects of BASIC on a Zenith Z-100 and an
IBM AT.  Gubbins@radc-tops claimed that the Z-100 with an 8 Mhz 8088 is
faster than an AT because of the one wait state penalty on memory reference
instructions on the AT.

  I don't want to drag that discussion to this newsgroup but not being
familiar with 80286 internals I was wondering if anyone knows precisely what
hardware in the Intel 286 makes its instructions take fewer cycles than on
the 8088.  Is it replacement of microcode with more hardware?  Pipelining?
Larger Cache?  Which instruction types are the 80286 best at and which are
they worst at?  (in comparision to the 8088 if possible) What documents from
Intel or Independent parties are available which describe in detail the
design of the 80286. How much CAD was used in the design of the 80286 and
how many man-years did it take?

Sam Chin
USENET: allegra!cmcl2!acf4!tsc2597
ARPA  : tsc2597.acf4@nyu