Branchless calculation of ref_offset.
authorMichael Niedermayer <michaelni@gmx.at>
Mon, 8 Feb 2010 03:22:17 +0000 (03:22 +0000)
committerMichael Niedermayer <michaelni@gmx.at>
Mon, 8 Feb 2010 03:22:17 +0000 (03:22 +0000)
commit542d4a756a260f86eedba1df225ed0b24ab55007
tree5d77a64a26139dda8b1ad7486c6294a9d33ec2f8
parentfb70c63f7091070ddd9d6adddfe12d8908561103
Branchless calculation of ref_offset.
7 cpu cycles faster.

Originally committed as revision 21685 to svn://svn.ffmpeg.org/ffmpeg/trunk
libavcodec/h264_direct.c