| Unregistered | 25th February 2004 04:35 |
I suppose a certain amount of "burn-in" could be concievably helpful in lowering core temperatures (one symptom commonly mentioned by adherents of burn-in) with thermal interface material that only becomes optimally conductive after it has passed a specific temperature point.
Those who used a bit too much thermal compound for instance, could have the excess more evenly arranged on the surface of the die as a result of reduced viscosity concomitant with increased temperature. Slight expansion of the thermal compound followed by reduced viscosity at temperature could result in some compounnd being squeezed off of the die, the heatspreader, or even a processor hotspot, resulting in something closer to a monolayer.
Certainly the traces in the chip itself are not becoming more conductive as a function of being overvolted, but there are a considerable amount of variables in the system outside of the processor core itself, which undeniably is not going to get permanently better as a result of additional applied voltage. | 