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