Having it working fine and then suddenly stop working is the typical failure mechanism for a CPU. All clocked semiconductor products (memory, CPU's, basically any chip whose sequence is directly controlled by a clock-like signal) should commonly fail like this when they fail - which is pretty rare. A process known as burn-in - which the 'real' burn-in performed by the manufacturer under carefully controlledd conditions (not to be confused with "home-brew" burn-in) - attempts to weed out parts that are likely to fail quickly, but under all processes a certain very small amount are likely to slip through this process. When they die, they tend not to die gradually (in fact, I can't really think of a way that it could die gradually) like a mechanical part. They either work, or they don't and when they die it should be suddenly and unexpectedly.
pkomma, you are wise to check your motherboard for damage too since CPU's that fail often take the power regulators with them with they go (less often the reverse is also true; the regulator dies taking the CPU with it). CPU's in the field tend to last a long time and generally it tends to be user error that is the culprit (not that I'm in any way saying this happened here... just that in my experience that was the case). Intel strives to keep it's failure rate extremely low, and I'm sure that AMD does likewise. I personally just think that this is probably just a case of bad luck.
Patrick Mahoney
Microprocessor Design
Intel Corp.
