A survey on thermal management techniques of microprocessors

Abstract

The basic theory behind cooling a processor is the concept that heat flows from a region having a low
temperature to a region having an higher temperature. The equation governing this transfer is Fourier’s law which can be
basically stated as q = k A dT / s, where A = area available for heat transfer , k = the thermal conductivity of the material ,
dT = temperature difference across the material, and s = material thickness. All the techniques that is discussed in the paper
uses this principle as the basics in one form or the other. Conventionally this has been done with the help of heat pipes and
an heat sink. Almost all of the CPU cooling done in consumer PCs and laptops follows this variation of the above
technology. Another method is using nanotubes in which, Joule’s law states that heat is produced by an electric current
flowing through a conductor and is directly proportional to the resistance and the amount of time, current flows in the
conductor. It can concur that the amount of heat generated is directly proportional to the number of transistors.
From this we can see that overheating in a microprocessor is a major issue and must be addressed right away.