[FFmpeg-devel] [PATCH] SSE-optimized vector_clipf()

[Vitor Sessak]

Loren Merritt wrote:
> This version is faster than x87 for me, though of course slower than 
> non-simd sse:

Thanks for the tip, it is indeed faster. New patch attached (with the 
vector_clipf_c() function split in two to avoid strict aliasing breakage 
warnings.

-Vitor
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dsp_vector_clipf4.diff
Type: text/x-diff
Size: 5436 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20090822/1d8ebe35/attachment.diff>