I'm not sure the AMD architectures will have many performance benefits with this. I've been following along during the implementation, and a lot of stuff...