ts-7000
[Top] [All Lists]

Re: [ts-7000] Re: Newbie question before buying TS-7xxx

To:
Subject: Re: [ts-7000] Re: Newbie question before buying TS-7xxx
From: "Breton M. Saunders" <>
Date: Tue, 03 Aug 2010 09:53:39 +0100
> thanks for the reply.
>
> I found ct-ng very good. I went through a steep learning curve but it really 
> did not give me too many problems.
>
> However, when I tried applying Martin's patches I just got shit. It could not 
> add 1+1 and get the right answer. I probably did something wrong or the patch 
> set was incomplete.
>    
I had exactly the same problems in building an EABI+FP kernel.  I 
concluded that it wasn't worth my time diagnosing and fixing the problem.

> I saw somewhere he was asking for people to pay him to sort out problems with 
> the patch set. OK if you have a bank backing your project I guess.
>
>
> Maverick Crunch (TM): maverick as in selling fundamentally screwed silicon 
> and discontinuing the product before ever releasing a working version. Crunch 
> as in brick wall for all the mugs thinking they bought floating point capable 
> hardware.
>    
Its terrible.  Probably the worst "FPU" ever made.

 From reading the bugs there are some pretty serious ones in there - 
like an invalid operation leading to memory corruption.  (So much for 
security.)  Its also 11 cycles per operation - so its faster than 
software FP; but pretty poor overall.

> So I suppose the answer to OP is don't buy this if you think you need f.p. 
> for audio, but it may be OK performance with software emulation.
>    
The solution is to use fixed point.  libmad (madplay) is great for 
playing mp3s and takes between 20 and 30% of cpu load on this chip.  
There are equivalent integer only versions for speex and vorbis 
(Tremor).  Speex can encode voice in realtime on this chip using a 
slightly lower sample rate.

I have run all three of these codecs on a TS7400, and all of them are 
very efficient, leaving plenty of cpu time left over for application 
programs.  From memory, I believe that Tremor was slightly more 
efficient than mad.  Speex was the best, coming in at around 15% load; 
but the comparison is unfair since I only applied it to speech; allowing 
me to cut the sample rate and encoding data rate.

   -Brett



------------------------------------

Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/ts-7000/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/ts-7000/join
    (Yahoo! ID required)

<*> To change settings via email:
     
    

<*> To unsubscribe from this group, send an email to:
    

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/

<Prev in Thread] Current Thread [Next in Thread>
Admin

Disclaimer: Neither Andrew Taylor nor the University of NSW School of Computer and Engineering take any responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the birding-aus mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU