Thanks for updating my info majkp. I'm interested in the hardware side of this since I dabble with it myself also.
During these crashes, when the server was still running at 4,5GHz, did you have a temperature monitoring on?
I, myself, need to keep my eyes on the avg temps from time to time, solely because dust will lead to BSODs if given enough time. And everyone knows that no matter how well ventilated the system is it will collect dust over time. Plus no matter how well you dedust the system it will never be as clean as a new system, so avg temps are higher in older systems than in new ones. The rise can be caused also the thermal paste getting cracks/going old.
