What I hate about mine is when overclocked high (not really that high over 3.6GHz), I can run 2xP95 and loop 3Dmark to test stability. It will often run for a long or short time, while still running P95 to keep the load and temp on, then I would browse the net, and at random Internet Explorer or either Mozilla will crash. This happens when the heat sets in. Does anyone else have this happen?
Yesterday I put the EE back on air to see where the limits are and really let it heat up. It crashed at 3.6GHz at only 60C (IE and Mozilla crashed far before at about 50C). Testing at 3.47GHz, it crashed at around 70C and the same instability beforehand.
From what I understand, CPU unstability kicks in far before the chip is ready to crash, even if some other programs or stress testers are still running.
These chips can be overclocked to high freq, but all you are really getting is just some numbers, and no true stability, unless you can cool it hardcore (air is useless and even water isnt that great). Operating on the edge stability is often actually not far away from stock speed, making them all in all crappy overclockers. Thermal throttling kicks in on the Gallatin at real low temps too.
Not sure what inside of CPU is the achilles heal is, but I do have my own opinions of what might be the cause.
