Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!sun-barr!cs.utexas.edu!csd4.milw.wisc.edu!dogie.macc.wisc.edu!uwvax!tank!eecae!cps3xx!flynn
From: flynn@pixel.cps.msu.edu (Patrick J. Flynn)
Newsgroups: comp.sys.next
Subject: Crunching with the DSP (was Re: Backups, ps4014, TeX, and a DSP...)
Message-ID: <3379@cps3xx.UUCP>
Date: 13 Jun 89 01:28:53 GMT
References: <3793@tank.uchicago.edu>
Sender: usenet@cps3xx.UUCP
Distribution: na
Lines: 20

In article <3793@tank.uchicago.edu> phd_ivo@gsbacd.uchicago.edu writes:
>
>>1.0 will come with an Array Processing library that provides functions
>>to do some number crunching on the DSP. 
>
>Good news...That may also calm some of the complaints made in the earlier
>message that thet 68882 is slow at addition and multiplication. (Damn slow...)
>

I was led to believe that the DSP does fixed-point math. Am I incorrect?

How about a preview of the features in the 1.0 array processing library?
Matrix multiplication? Inversion? Eigenanalysis? Decompositions?
(Is LINPACK being implemented on the DSP ;-))?  FFTs and spectral
analysis seem like a natural application for the chip.

------------------------------------------------------------------------------
Pat Flynn, CS, Mich. State U | "Don't eat with your hands, son; use your
flynn@cps.msu.edu            |  entrenching tool." 
(517) 353-4638               |                       -Firesign Theatre