But why would simply running a CPU at a higher clock speed cause it to interpret logic highs as logic lows? To answer that, we need to discuss some basic electrical principles.
In an ideal electronic device, your voltage would look like this. It would fluctuate between VSS
, but always be either a perfect
, or a perfect
. In this case, our interpretation would be prefect. Logic low and logic high would always be easily differentiable.
However, we all know ideal models usually aren’t exactly true. In fact, in a real electronic device, the voltage would actually look more like this. Rather than going from exactly VSS
, a small amount of time is necessary for the voltage to transition from one voltage to another. This is aptly termed the ‘transient response’.
One way to think of it is like a door. A door can be either opened or closed, yes, but can it ever go from opened to closed instantaneously? Of course not. There’s always some small amount of time required for the door to transition from state to state. The voltage in an interconnect is exactly the same; some very small amount of time is always required to transition from one voltage to another, and during that time, the voltage lies somewhere in between (when you close a door, for an instant, it’s somewhere between opened and closed).
When we try to overclock, this becomes a bit of a problem. Overclocking is increasing the speed, or frequency, at which the device operates. Increasing the frequency means decreasing the time between each cycle. If we decrease the time between cycles enough, we reach a point at which the cycle time is less
than the transient time. In other words, the cycles are so short that the voltage doesn’t have time to reach the levels it should. Consider the image below.
Effectively, we’re cutting off the voltage before it has a chance to reach its intended value. If we cut it off soon enough, we’ll see a voltage below the tolerance of the transistor, and all of a sudden, our logic highs become logic lows. Therein lies the problem.
If you’re having trouble grasping that, there’s a little experiment you can try on your own, using the door example we gave earlier. Go stand in a doorway, and try to open and close the door completely at a frequency of 1Hz, or once per second. Pretty easy? Okay, now try to open and close the door five
times per second. Can’t do it, can you? What you ended up doing is likely flipping the door between your hands, perhaps a foot apart, five times per second. You had to do it so fast that you didn’t have time to let the door completely open or close. It never reached its intended state -- it simply couldn’t turn about and reverse direction and position that fast. Give it a try.