It's like candy bars. They keep making them smaller at the same price to increase profit.
Seriosly though, there are good reasons for smaller chips. The first and foremost is that more chips can be fit onto a wafer. For large chips, the cost is all about what fraction of a wafer it uses. The cost to process a wafer is pretty much fixed, regardless of how many chips result from it.
Using less of the expensive wafer is only one part though. Yield is the other. All wafers have imperfections. Think of them as being small but randomly scattered about the wafer, and any IC that hits one of these imperfections is trash. When the wafer is covered by lots of small ICs, only a small fraction of the total are trash. As the IC size goes up the fraction of them that hit a imperfection goes up. As a unreal example that nonetheless points out the issue, consider the case where every wafer has one imperfection and is covered by one IC. The yield would be 0. If it were covered by 100 ICs, the yield would be 99%.
There's a lot more to yield than this, and this is greatly oversimplifying the issue, but these two effects do push towards smaller chips being more economical.
For really simple ICs, the packaging and testing cost dominates. In those cases, the features size is not so much a driving issue. This is also one reason we have seen a explosion of smaller and cheaper packages lately. Note that extreme small features size is being pushed by very large ICs, like main processors and GPUs.
In general:
Take whatever physical system to an extreme, and all the simple models which were developed by engineers will break apart.
Simple model for active power dissipation:
The statement about an exponential increase in heat dissipation at extreme overclocking is not consistent with the following equation:
$$P_g \propto C_gV^2f$$
but how the above equation was derived?
Well, it is based on the following simplification:
simulate this circuit – Schematic created using CircuitLab
This model assumes that:
- Transistors behave like an ideal, mutually exclusive switches (no overlap in time when both switches are ON)
- All capacitances may be represented as a single equivalent capacitor at the output
- No leakage currents
- No inductances
- More assumptions
Under the above assumptions, you can think of inverter's (or any other logic gate's) action as of charging the output capacitor to \$V_{dd}\$ (which consumes \$\frac{1}{2}C_{tot}V_{dd}^2\$ Watt from the power supply), and then discharging it to ground (which does not consume additional power). The frequency factor \$f\$ is added to represent an amount of such cycles per second.
In fact, it is surprising that the above equation may be an accurate estimation of dynamic power at all, given the huge amount of non-trivial assumptions made. And indeed, this result may be used for the first order analysis only - any serious discussion of power dissipation in modern CPUs can't rely on such a simplified model.
How the simple model breaks:
All the assumptions made while developing the above simplified model break at some point. However, the most delicate assumption which can't hold for an extreme frequencies is that of two mutually exclusive ideal switches.
The real inverter has non-ideal Voltage Transfer Curve (VTC) - a relation between inverter's input and output voltages:
On the above VTC the operational modes of both NMOS and PMOS were marked. We can see that during switching there will be time when both NMOS and PMOS are conducting at the same time. This means that not all the current drawn from the power supply will flow to "output capacitor" - part of the current will flow directly to ground, thus increasing the power consumption:
What this has to do with frequency:
When the frequency is relatively low, the switching time of the inverter comprises negligible part of the total operational time:
However, when the frequency is pushed to the limit, the inverter "switches continuously" - it is almost always in switching activity, thus dissipating a lot of power due to direct ground path for the current (time scale changed):
Maybe it is possible to try to model this and see if the result is exponential, but I prefer to use simulations (however, the simulation will account for all non-idealities, not just this one).
Simulation results:
In simulation I measured the total energy (integral of power) drawn from an ideal power supply by an inverter in the following configuration:
The first and the last inverters are there just in order to model a real driving and loading conditions.
The dissipated energy as a function of frequency:
We can see an approximately linear dependence for periods longer than 1ns, and clearly exponential dependence for shorter periods.
Notes:
- For the simulation I used an antique 0.25um transistor models. The current state of the art transistors are more than x10 shorter - I guess the divergence from the linear model is stronger is newer technologies.
- The question whether a particular CPU/GPU can be overclocked such that it enters the exponential frequency dependence state while still stable and functional is device specific. In fact, it is exactly what overclockers try to derive empirically - to what frequency can a given device be pushed without malfunctioning.
- All the above results and discussions do not consider changing voltage levels. I guess there is no way to analytically predict the outcome of simultaneous change of both frequency and voltage - the only way to find out is to perform an experiment.
From a single inverter to CPU:
CPUs mainly consist of logic gates, which are conceptually similar to an inverter. However each modern CPU has sophisticated measures of controlling its operating frequency, operating voltage and can turn off its submodules during runtime. This means that the heat dissipation trend of the whole processor may be slightly different than this of the single inverter. I guess that the statement about exponential increase in heat dissipation during extreme overclocking is a bit of exaggeration, but we are not mathematicians: either it is exponential, or \$\propto f^{3+}\$ - it is all kind of "bad".
Best Answer
They aren't cooled from below because they have pins on the bottom, and FR4 below that.
Due to having a much lower thermal conductivity, $$ \begin{array}{rrl} \text{Copper:} & 385\phantom{.25} & \frac{\mathrm{W}}{\mathrm{m}{\cdot}\mathrm{K}} \\ \text{Aluminum:} & 205\phantom{.25} & \frac{\mathrm{W}}{\mathrm{m}{\cdot}\mathrm{K}} \\ \text{FR4:} & 0.25 & \frac{\mathrm{W}}{\mathrm{m}{\cdot}\mathrm{K}} \\ \end{array} $$the material on the bottom of the CPU would transport much less heat.
Odds are you wouldn't want to surround the signals with metal which would change the impedance drastically, so metal on the bottom is more of an issue. If you did build a socket out of metal, it would need to be micromachined, which would be many times more expensive than a plastic injection molded socket. These things would prevent you building a processor socket that would wick heat away.
You could put a cooling block on the bottom of the board, but the PCB material (FR4) would reduce the cooling substantially.