[Ffmpeg-devel] Fixed point arithmetic RealAudio G2 (cook) decoder

Guillaume Poirier gpoirier
Thu Mar 8 16:08:52 CET 2007


Luca Barbato wrote:
> Ian Braithwaite wrote:
> 
>>Hi,
>>
>>
>>OK - here's my fixed point patch for the cook audio decoder.
>>
> 
> 
> Looks interesting, maybe it's a starting point to have a faster cook
> using vector instructions where available.

Since when does vectorization only works on fixed point arithmetic?

Major SIMD instruction sets (Altivec, SSE, 3Dnow!) all support
floating point SIMD. Only MMX is restricted to integer.

BTW, BLAS-type computation is probably the easiest and most
straightforward candidate for vectorization (since you don't need to
pack/unpack).

Or maybe I missed smth in your reasoning?

Guillaume




More information about the ffmpeg-devel mailing list