Xref: utzoo comp.sys.next:18193 comp.dsp:1727
Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!sdd.hp.com!news.cs.indiana.edu!ariel.unm.edu!cs.umn.edu!ux.acs.umn.edu!vz.acs.umn.edu!pukite
From: pukite@vz.acs.umn.edu (J. PUKITE)
Newsgroups: comp.sys.next,comp.dsp
Subject: Re: DSP Chip in NeXT -- how can we use it?
Message-ID: <4023@ux.acs.umn.edu>
Date: 30 May 91 02:38:16 GMT
References: <1991May25.023910.8736@shaman.com>
Sender: news@ux.acs.umn.edu
Reply-To: pukite@vz.acs.umn.edu
Followup-To: comp.sys.next
Organization: DAINA
Lines: 21
News-Software: VAX/VMS VNEWS 1.3-4

In article <1991May25.023910.8736@shaman.com>, jiro@shaman.com (Jiro Nakamura) writes...

> [... deleted]
>	I'd especially like to hear from academia. What are the scientific
>uses of the DSP? Or is it an albatross? I think that a NeXT without the DSP
>is like castrating it into a Mac, but what can we use it for? Can we
>explore some possibilities? How about the ten wackiest things to do with
>the DSP (no boring things like: fax modem, ECG recorder, powerglove monitor,

We have investigated some scientific applications using a
floating-point DSP in statistical computations and simulation.
Using an AT&T DSP32 we were able to achieve an average
computation speedup of 100 over a conventional 20 MHz 386.

For details see our paper in the Winter 1990/1991 issue
of Simulation Digest (a joint publication of ACM SIGSIM and
IEEE Computer Society TCSIM): "Digital Signal Processors
for Computation Intensive Statistics and Simulation", pp.20-29.

J.Pukite/DAINA
pukite@vz.acs.umn.edu