[FFmpeg-devel] MMX accelerated DSP functions for VC1/WMV3 decoders

Michael Niedermayer michaelni
Sat Jun 30 22:43:14 CEST 2007


On Sat, Jun 30, 2007 at 10:13:29PM +0200, Christophe GISQUET wrote:
> Good evening,
> Michael Niedermayer a ?crit :
> > no, you can design a cpu where its not
> That was not my question. Your suggestion shows that you know
> instruction cycle counts or pairing restrictions. I don't. Hence I would
> have liked to know if that suggestion was based on one experience or not.

i didnt check all docs of all currently existing cpus, i also didnt benchmark
it, still paddw is faster then psllw for what i remember maybe you should
RTFM if you want to know it more precissely
(see http://www.agner.org/optimize/)

also paddsw is not equivalent to psllw, paddw is

Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

I am the wisest man alive, for I know one thing, and that is that I know
nothing. -- Socrates
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20070630/85a948ee/attachment.pgp>

More information about the ffmpeg-devel mailing list