Agreed--
The point being though that we are a good 70C or so away from needing to worry about catastrophic thermal effects. If the chip gets too hot, critical speed paths start to run too slow, which would result in a functional failure until the device cooled back off. Throttling assures that there will not be such a functional failure, let alone any other sort of thermal runaway. Ever since the first few faulty analysis of Willamette, I have not heard of any P4 design entering throttling when used as specified.
--Alan