[vorbis] New decode chip

[Shawn]

The fact that gcc 3.1.1 produces 10% faster code on my athlon than 2.96
would seem to suggest that there is room for improvement in the way (at
lower levels) things are done.

The truth is, finding the best way can be done by evolving code. You
just have to think of the algorithm at it's lowest levels right down to
the nuts and bolts.

You are correct to say that this does not fully exploit the power of
genetic algorithms.

On 08/06, fungus said something like:
> Shawn wrote:
> > A functionality must have a certain level of specificity in order to
> > evolve a solution.
> > 
> 
> This thread is total rubbish, you can't "evolve" things
> like ogg decoders any more than you can "evolve" an
> algorithm to add two numbers. Adding numbers is a
> precisely defined algorithm with no fuzziness, you
> can't vary the algorithm a bit to see if the numbers
> add together "better".
> 
> Genetic algorithms are used for tasks where there's a lot
> of fuzziness and the "correct" answer would otherwise be
> found by trial and error, eg. creating the shape of a
> racing boat hull.


--
Shawn Leas
core at enodev.com

I went to a general store.  They  wouldn't  let  me  buy  anything
specifically.
						-- Stephen Wright

<p>--- >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.