Implement speed optimizations in x264 encoding.

First step would be a lot of profiling (not only for cycles spent,
also for cache misses etc.).

Second step would be optimizing code that presents possibilities for
doing so.

Third step would be testing algorithmic optimizations if any present
themselves (or other people suggest them - i am more than open for
any input the project has on this)

This does not need to be limited to x264. Anything you want to be done in this regard for VLC can be done under this proposal too. As long as its speed optimization and simd related, i'm all for it.