[rescue] Perverse Question

[Joshua D. Boyd]

On Tue, Jun 10, 2003 at 11:46:38PM +0200, Frank Van Damme wrote:
> On Tuesday 10 June 2003 20:49, Joshua D. Boyd wrote:
> > Oh for crying out loud.  There have been assembly optimized MP3 decoders
> > for a lot longer than there have been MMX chips.  An old one free for
> > non-commercial use that comes to mind is amp (which was once used as the
> > decoding engine for winamp, and also was used for numerous command line
> > programs).
> >
> > There have also been realtime mp3 decoders for lots of older, low clock
> > rate Motorola platforms, like the Amiga, the Next machines, and probably
> > Mac OS.
> 
> Perfectly possible. I don't know crap about assembly (or even programming :)
> ), the only thing I said was that x86 cpu's like the pentium mmx and better
> have instructions made specifically for multimedia.

No, you said:
On Tuesday 10 June 2003 15:01, Joshua D. Boyd wrote:
>> Also, many MP3 programs use hand optimized intel assembly (or they
>> used to back when I was doing Mp3s on a 486).  It might be that such
>> a thing is needed to make the sparcs work well.  Finally, it might be
>> that Intel has an instruction that helps that sparcs don't.

> They do, but not an 486 :-)

> Think mmx...

which to any reasonable person means that you are denying my claims
about optimization for 486 level machines.
 
> btw, I find winamp pretty cpu-hungry...

For many years now, Winamp used the reference decoder from that German
institute, and I expect that they still do.  It is straight C to my
recall.  I suppose they could have switched decoders again since I last
heard.  It was now many years ago that WinAmp could actually be called
fast.