Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83 based; site houxj.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!houxj!wapd
From: wapd@houxj.UUCP (Bill Dietrich)
Newsgroups: net.arch,net.lsi
Subject: Re: AT&T digital signal processor with floating point
Message-ID: <535@houxj.UUCP>
Date: Thu, 11-Jul-85 10:00:37 EDT
Article-I.D.: houxj.535
Posted: Thu Jul 11 10:00:37 1985
Date-Received: Sat, 13-Jul-85 07:47:11 EDT
References: <813@turtlevax.UUCP>
Organization: AT&T Information Systems, Holmdel NJ
Lines: 23
Xref: watmath net.arch:1540 net.lsi:114


The programmable digital signal processor is called (internally,
at least) the DSP-32.  I don't know much about it, but I know
it has been available for some time and I'm pretty sure it
is for sale on the commercial market.  I don't know the part number
or where to call for data sheets.

My understanding is that the DSP-32 functions as a high-speed
part in a dedicated system.  In other words, it is not appropriate
to feed it from a general-purpose microprocessor.  In contrast,
the IEEE floating point part you mentioned IS part of a general-
purpose chip-set, the WE 32100 chip-set.

This IEEE-standard (draft 10) part is the WE 32106 Math Acceleration
Unit (MAU).  It has a special protocol to work with the WE 32100
CPU, but also has a peripheral mode so that one could hook
it to other commercial CPUs.  It is not microprogrammable.
To get a data sheet, call  (800) 372-2447.


					Bill Dietrich
					AT&T Information Systems
					houxj!wapd