[FFmpeg-devel] [PATCH] More H.264 decoding speed tweaks

[Jason Garrett-Glaser]

> (decoding times in seconds, stddev of 10 runs)
> 1280x528 qp12 10153kbit  svn: 135.70 +/- .13  patched: 137.65 +/- .15
> 1280x528 qp18  3672kbit  svn: 104.53 +/- .12  patched: 107.37 +/- .15
> 1280x528 qp24  1408kbit  svn:  68.98 +/- .09  patched:  72.50 +/- .10
> 1280x528 qp30   653kbit  svn:  54.68 +/- .07  patched:  58.57 +/- .07

Odd that my benchmarks were otherwise; perhaps its more
source-dependent than I thought?

One question, Loren, while we're on this topic; how should your SSE2
iDCT4x4 be implemented when we start bringing x264's nasm asm to
ffmpeg?  Should we check the DC coefficient for each 8x4 block and
make an 8x4 dc_add, or what?

Dark Shikari