"look up the words "context" and "thread" and report back."
You must be one talented kid if you can express your brilliant ideas in such a short form. Not!
Here is your other context: "I claim that die temperature increases with power *density*, even if power is constant, assuming everything else (ambient temp, heatsink) is the same."
First, your claim is obvious and trivial. The problem of heat transfer from (essentially single-point-source) semiconductor devices (mostly power FETs, rectifiers, etc.) has been known to electronic industry for decades. Yes, if the die becomes smaller, the effective thermal resistance of heat spreader becomes slightly higher, and the die temperature will rise proportionally.
Now, please note the term "die" above.
What you obviously fail to comprehend is that your claim has no relevance with the reported observation. You miss the main thingy: the temperature reading device in the HKEPC report was the "CoolerMaster AeroGate 3". It uses its own sensors for temperature readout. Therefore, it measured temperatures somewhere at the HEAT SINK at most, at some place that is likely not even close to the center of heat spreader interface. In any case, your brilliant theory, although theoretically correct wrt die temperature, does not apply to the observed readings. Now engage all your "mere university degree" and admit that your straw arguments to justify humongus 67C temperature of all-copper HEAT SINK on a 90nm processor are pathetic.
Hence, go lookup "ignoramus", and please do not report back, you'll look not very good.
- Ali
BTW, congratulations I guess, the single vendor of "Athlon 64 90nm" processors (monarchcomputer.com) now lists all three grades as "in stock". Go celebrate. |
