Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!sdd.hp.com!think.com!yale!quasi-eli!cs.yale.edu!news
From: kuszewsk@euler.biology.yale.edu (John Kuszewski)
Newsgroups: comp.sys.next
Subject: Re: new machine
Message-ID: <1991Jun13.162641.10062@cs.yale.edu>
Date: 13 Jun 91 16:26:41 GMT
References: <1991Jun12.200749.19703@data.com>
Sender: news@cs.yale.edu (Usenet News)
Reply-To: kuszewsk@euler.biology.yale.edu
Organization: Yale University, Department of Computer Science, New Haven, CT
Lines: 23
Nntp-Posting-Host: volta.biology.yale.edu

Steven M. Boker writes
> In article <1991Jun12.183133.28899@cs.yale.edu>  
kuszewsk@euler.biology.yale.edu writes:
> >Hi!
> >
> >Does anyone have a feel for the 88110's double precision FP speed?  Will it  
be  
> >tripled as well as the integer speed? (please?) If it is, we're talking  
about  
> >some _serious_ processing--about the level of a Stellar GS2000 or SGI 4D/GTX  
> >series.  
> >
> 
> The 96002 is a floating point DSP.  Matrix multiplies that will make
> your eyes swim
> 
> Steve

The 96002 is a _single_precision_ floating point DSP that can do 50 MFLOPS at  
33 MHz.  Double precision is required for some applications but cannot be done  
with a 96002.  Plus, will it be available for running jobs?  The ND's 860  
isn't, partially because it's alkready busy running the screen.  I'm wondering  
what's actually available for user jobs.