Just curious

[WhatMeWorry]

I've never overclocked and probably never will, but when one overclocks, is the power required/heat produced pretty much a linear function of the overclock amount?

[Flying Fox]

No. It can even go thermonuclear if you push things high up a lot. That is why you see those liquid nitrogen stuff used in overclocking competitions/exhibitions/demos such as this.

[just brew it!]

This is mainly because of the increased core voltage which typically accompanies any serious overclocking attempt. Power consumed by a CMOS circuit is a fairly linear function of frequency at a given voltage; but power consumption (and heat dissipation) increases according to the square of the supply voltage.