In an effort to get back extra CPU cycles for accuracy, I've taken the SSE/vectorized plugin from my simulator and ported the changes into FatCat's plugin. Note that it does require your CPU to have the SSE4.1 instruction set. Intel CPUs have had this for a few years now (since Penryn?) while AMD users are out of luck unless they have Bulldozer (I believe).

 

The performance difference is incredible. I didn't think SSE would make that much of an impact, but I was completely wrong. I went from not being able to even come close to playing Conker's BFD with a LLE RDP plugin to a completely smooth gameplay experience. CPU load was previously 100%, with video and audio stuttering. CPU load is now sitting at 88% in the main lobby of Conker's BFD; no audio skips and crystal clear audio.