Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site l5.uucp
Path: utzoo!watmath!clyde!bonnie!akgua!gatech!ut-sally!seismo!lll-crg!l5!gnu
From: gnu@l5.uucp (John Gilmore)
Newsgroups: net.arch
Subject: Re: XT/370 is microcoded 68000s
Message-ID: <330@l5.uucp>
Date: Wed, 11-Dec-85 06:27:33 EST
Article-I.D.: l5.330
Posted: Wed Dec 11 06:27:33 1985
Date-Received: Fri, 13-Dec-85 07:46:31 EST
References: <632@mit-eddie.UUCP> <2191@gatech.CSNET> <373@amiga.amiga.UUCP> <2207@gatech.CSNET>
Organization: Nebula Consultants in San Francisco
Lines: 23

In article <2207@gatech.CSNET>, noman@gatech.CSNET (Jeffrey L. Grover) writes:
> In article <373@amiga.amiga.UUCP>, bruceb@amiga.UUCP (Bruce Barrett) writes:
> > 	Well, IBM has had practice with this sort of thing...The IBM
> > XT/370 emulated a 370 by re-microcoding 1 (or 2??) 68000's and an 8087
> > (for floating point).
> 	yep, IBM put 2 68000's in the XT/360
> 		1. microcoded to execute as a 360
> 			sans floating point operations
> 		2. microcoded as a floating point
> 			co-processor

That's not quite true.  One 68000 was re-microcoded to run most of the
360 instruction set.  The other 68000 was a straight-out-of-the-box
68000, running 68000 assembler language, that implemented the less common
360 instructions by emulation.  A third chip was an Intel 8087 specially
modified to run IBM-format floating point rather than IEEE float.

At first a lot of people thought there was great promise in the XT/370,
but virtually nobody outside IBM data centers has bought them, because
the software is exhorbitantly expensive, and it comes on 9-track magtapes.
(What does Assembler H rent for, per month, these days?  Certainly 
hundreds of dollars...)  If you've bought or rented the software for your
mainframe, you can sublicense it for your XT's; but if you haven't...