[vorbis-dev] Optimisations

[Segher Boessenkool]

> That said, I presume fixed point would have to be the choice for x86
> optimisations, as MMX is the only common SIMD instruction set for all post
> 1997 x86 CPUs.  Implementation in SSE or 3DNow! would be redundant.(?)  I
> am just looking for an excuse to learn and apply MMX/3DNow (I have a Duron)
> in some sort of constructive fashion. :)

Actually, it makes sense to have separate Intel/AMD optimizations anyway,
as both families of processors benefit from very different optimizations.
The Athlon likes straight, no-fuss code, with the occasional carefully
placed prefetch instructions (they help a _lot_! Especially the
"this cacheline will be dirtied" and the "no L1/no L2/etc." variants),
while the P-2/III likes very processor dependant optimizations (which hurt
the K6 and other, but Athlon is reasonably immune). I suggest we take
the most
time-critical parts, and optimize-em to hell, _per processor_. Maybe we can
make some autotuning version (choosing different implementations on the run).
I'm investigating, would like input from others, of course :-)

> Still, is some sort of optimised (platform dependant optimisations where
> available) DSP library that both the encoder and decoder can use worth it
> at this stage?

It would help for all the usual dsp ops (vector copy, interleaved copy,
fft, etc.) But I wouldn't write one myself, good ones are available for all
serious processors.

Ciao,

Segher

--- >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-dev-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.