Let’s face it – we’re a size, speed, and power obsessed society. SUVs with three hundred horses, 53” TV screens, and 3+ gigahertz (Ghz) computer processors. Every level of size, speed and power has a trade-off. The 300 HP SUV probably displays its fuel efficiency in “gallons per mile”. The 53” TV screen may be huge, but have an inferior picture compared to its smaller siblings. And, the 3 Ghz computer processor uses a lot of power and produces enough heat to double as a George Foreman grill in its spare time. Over the next two weeks we’re going to take a layman’s look at computer processors, concentrating on the most commonly used (and misused) measure of their power – clock speed.

Notice I haven’t talked about computer “chips”. The modern computer processor is really made up of a several parts that work together, each part contributing to the speed at which work gets done. We’ll refer to this “team” as the computer’s Central Processing Unit, or CPU. The CPU consists, basically, of the processor, data routing conduits known as “busses” (from the electrical term, not the big yellow variety), and data staging areas known as caches. Trying to keep it simple, when your computer is computing, it is getting data from some input device (your keyboard, your hard drive, etc.), bringing it into the CPU, where it may be staged in a cache before it is delivered to the processor by a bus. The processor then does its work, and the data is then sent back out a bus into memory, or to an output device, like your screen or printer. Obviously, the faster the processor or bus, or the larger the cache, the faster and more efficient the processor can become.

Inside the processor itself, data enters what we’ll envision as a pipeline, with different stages where data is fetched, decoded, executed and stored. Let’s say you’re using your computer’s digital photo program to make a picture smaller to attach it to an email. As your program tells the processor to make the image smaller, the picture data flows into the pipeline, is worked on, and flows out of the pipeline, eventually showing up on your screen as a smaller picture. More on the pipeline as we go along.

Windows-based consumer computers use either Pentium (Intel) or Athlon (AMD) CPUs. Both are descendants of 8086 chips (we’ll call them x86) first introduced in 1978. x86 processors are known as CISC (Complex Instruction Set Computing) processors. As the name implies, CISC CPUs have highly complex processing instructions, and, using the pipeline metaphor, have a relatively narrow and long pipeline.

Macintosh computers use PowerPC chips, supplied by both Motorola and IBM. The PowerPC chip was born in 1993, and is known as a RISC (Reduced Instruction Set Computing) chip. I know it’s getting “geeky”, but hang with me. RISC CPUs have (you guessed it) reduced and simpler processing instruction sets, and, metaphorically, wider and shorter pipelines.

The clock speed of a processor is basically how many times the CPU “ticks” per second. Every time it “ticks”, an instruction on a piece of data gets performed. Consumer Pentium 4 and Athlon x86 processors are running in the 2.5 GHz range, and high-end siblings are topping out at clock speeds of over 3 GHz – that’s 3 billion ticks per second.

PowerPC “G4” processors in Macintosh computers top out at 1.4 GHz, yet provide performance equal to or better than their Pentium counterparts having twice the clock speed. How can that be? Here’s a hint – it’s all about the pipeline. Tune in next time!

© 2003 Peter F. Zimowski