diff x86/mmx.h @ 11042:74f672a1f763 libavcodec

Change wraper functions to always inline, they are faster now that way. 1% faster MBAFF decoding overall, maybe ~0.1% faster for the cathedral sample.
author michael
date Thu, 28 Jan 2010 11:37:35 +0000
parents 7768bdfd4f7b
children
line wrap: on
line diff