Newsgroups: comp.speech,comp.compression,comp.dsp
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!pipex!uunet!news.crd.ge.com!usenet
From: koch@reggae.crd.ge.com (Steven R Koch)
Subject: Re: PC Binaries for CELP (FS1016) ?
In-Reply-To: noelg@jtec.com.au's message of Wed, 7 Dec 1994 15:50:31 EST
Message-ID: <KOCH.94Dec12112719@reggae.crd.ge.com>
Lines: 24
Sender: usenet@crdnns.crd.ge.com (USENET News System)
Nntp-Posting-Host: reggae.crd.ge.com
Organization: GE Corporate Research & Development
References: <3bnf89$f66@monet.ecs.soton.ac.uk> <3bvmeb$cps@colossus.holonet.net>
	<noelg.57.00C0D7B4@jtec.com.au>
Date: Mon, 12 Dec 1994 16:27:18 GMT
Xref: glinda.oz.cs.cmu.edu comp.speech:4035 comp.compression:16096 comp.dsp:15205


>>It runs 14 times real time on a 90Mhz Pentium (although you don't always
>>get exactly the same results!)

Must be the Pentium fdiv bug! :-)

>    Damn, FS1016 runs real-time on my 50 Mhz DSP32c. Goes to show a 50 Mhz
>    DSP32c will suck the doors off a 90 Mhz Pentium anyday (and give the same
>    results each time).
> 
>    noel gordon
>    noelg@jtec.com.au

Is everyone running the same FS1016 C code?  (Assembly language can speed
this stuff up tremendously.)  If so, I'm impressed; otherwise, its not a
meaningful comparison.

-- 
--
Steven R. Koch        			kochsr@crd.GE.COM
GE Corporate R&D      
Schenectady, NY       			(518) 387-5217

"I like maxims that don't encourage behaviour modification." -- Calvin
