Newb Question: Heat produced by CPU is the same in spite of its fan, right?

[SaltBoy]

Suppose I have a CPU that at full load stays at 55C. If I replace its fan with a lower-speed/lower-noise fan, that won't actually change the amount of heat output by the CPU, it will only change the CPU temperature. Right?

 

[Bull Dog]

Actually CPU's tend to use a little more power the hotter they get. But its a fairly insignificent amount.

 

[Maximilian]

Yea, heat outputs the same at full load as long as voltage & clockspeed are constant. The temp will be higher as the heatsink will dissipate heat less efficiently due to the lower fan rpm. However the cpu still outputs the same amount of heat as it did before.

 

[Zolty]

Originally posted by: Soviet
Yea, heat outputs the same at full load as long as voltage & clockspeed are constant. The temp will be higher as the heatsink will dissipate heat less efficiently due to the lower fan rpm. However the cpu still outputs the same amount of heat as it did before.

yes

 

[Some1ne]

Yes, liken it to sticking a 100W light-bulb in the freezer vs. sticking the same bulb in the refrigerator. In either case, it still generates 100W of heat, but the freezer can dissipate that more effectively, so the freezer bulb will be cooler to the touch.

 

[nrb]

In either case, it still generates 100W of heat, but the freezer can dissipate that more effectively, so the freezer bulb will be cooler to the touch.

When you're talking about CPUs and freezer temperatures, this actually isn't true. I used to own a Kryotech Super G Plus system, which was a 750MHz Slot-A Athlon clocked to 1 GHz. This was achieved by cooling the CPU to about -40 degrees C. Kryotech referred to this not as "over-clocking" but as "thermal acceleration" and (for once!) this was not purely a marketing term. When you cool semiconductor chips to that sort of temperature, the amount of power they use actually significantly diminishes. So the amount of power the chip in that system used when running at 1 GHz was actually almost the same as the power it would have used running at 750MHz at normal temperatures. This meant that (for example) there was no shortening of the chip's lifespan in the way there is with overclocking.

However, if you're talking about differences of a few degrees C in temperature then yes, the difference in power usage is negligible.

 

[Some1ne]

When you're talking about CPUs and freezer temperatures, this actually isn't true. I used to own a Kryotech Super G Plus system, which was a 750MHz Slot-A Athlon clocked to 1 GHz. This was achieved by cooling the CPU to about -40 degrees C. Kryotech referred to this not as "over-clocking" but as "thermal acceleration" and (for once!) this was not purely a marketing term. When you cool semiconductor chips to that sort of temperature, the amount of power they use actually significantly diminishes.

I'd liken that more to putting the 100W bulb in a vat of liquid nitrogen...a household freezer wouldn't be sufficient to keep it that cold, I don't think. But anyways, yes, the colder you make something, the more efficiently it will conduct energy (to the point where in theory at absolute 0 you get a superconductor). But like you said, the temperatures necessary for this sort of thing to start to come into play are well below anything attainable with air cooling.