http://www.xbitlabs.com/articl...pu/display/replay.html' ">http://www.xbitlabs.com/articl...pu/display/replay.html
I've only skimmed it, but it's interesting to see how Intel's engineers tried to compensate for the P4's long pipeline with a feature that eats up countless cycles. It also shows why Intel decided to implement Hyperthreading.
-------------------------
If all else fails, hit it harder.