[FFmpeg-devel] [PATCH 2/2] VC1: MMX 8x8 IDCT.

Loren Merritt lorenm
Thu Feb 17 02:18:48 CET 2011


I would pull transpose and store out of the macro, and s/movq/mova/g, so 
that it can be instantiated for sse2.
Commit message should specify the speedup for this individual function, as 
well as for the codec as a whole.

--Loren Merritt



More information about the ffmpeg-devel mailing list