[vorbis] MMX/3dNow! etc

[Gian-Carlo Pascutto]

On Sat, 29 Sep 2001, safemode wrote:

> What you're saying is that it's not worth the effort.  But for the
> many people who use spare boxes for "music consoles" or those who just dont
> have strong hardware, it may make the difference of being able to play ogg
> files and being forced to stick to mp3s.
> I just think there should be some testing to see if such instructions would
> be worth taking the time to program before they're brushed away.

Please, where did I say it was not worth the effort or that
they 'should be brushed away' ?

The issue is that right now it is still possible to make the
vorbis encoder and decoder significantly faster _without_ _having_
to revert to using architecture-specific instructions.

It is of little use to recode large parts of vorbis using MMX/SSE/
3dnow if the original algorithms can still be speeded up a lot.

That does _not_ mean it won't be worthwhile to use them for
optimization in the future (at the time 1.0 comes out for example).

You can get a small speedup from those isns right now without
hand effort by recompiling the libs with Intel C 5, which can
convert things into MMX/SSE code automatically. But it is not as
efficient as a manual conversion of course.

If you want to give that a test, I can make a compile for you.

PS. My Cyrix120 was able to decode RC2 files at 75% cpu. Because
it's FPU is about as fast as a Pentium 75, I think that a Pentium
100 should have no problems decoding vorbis.


-- 
GCP

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.