Stand back, folks. This article’s going to be just a touch more technical than some of my other stuff.
So, here’s the deal. You see a new desktop processor on the market. It runs at about….2 GigaHertz. Now, those of you with any computing experience will know right off that 2 GigaHertz in a desktop is a pretty abysmal clock rate (the rate at which the system completes a single cycle). Obviously, that processor over there that runs at 2.6 GHz is better, right?
This is something that’s been circling the web for a while. It’s known as the MegaHertz (or GigaHertz) myth. No one’s quite sure how – or why – it started, though it’s likely that it came about because people saw the clock rate of a processor as a simple, no-fuss way to determine how fast or powerful that processor was. Overzealous tech bloggers didn’t do a great deal to debunk the myth, either. Nor did Intel, which has been pushing the “higher clock rate = better processor” deal for quite some time.
One CPU might seem better than another CPU simply because it completes twice as many cycles as another CPU (hence, it has a higher MHz/GHz value). However, the other CPU could very easily complete twice as much with each cycle, meaning that both CPUs would ultimately process the same amount of information.
Its not just the clock rate that determines how powerful a processor is. The microarchitecture of the processor also plays a HUGE role in its quality.
I’ll use an analogy that most anyone should be able to understand. Let’s say you have two factories- A and B. The factories each represent different computer processor designs. Now, let’s say the workers in factory A work for eight hours per day- we’ll say that equates to 2.4 GHz. The workers in factory B, on the other hand, only work for four hours per day- 1.2 GHz. Naturally, you’d expect the workers at factory A to produce more stuff than those at factory B.
Thing is, though, factory B has a far better assembly line- and harder working employees- than Factory A. As a result, they end up putting out the same volume of products as factory A, even though they only put in half the hours.
It’s not a great analogy, but it should at least give some idea as to why clock rate isn’t the be-all and end-all of measuring a processor’s raw power.