We often view computer technology as a science rife with startling discoveries and dramatic innovations. Certainly that's how it sounds when you read the press releases from the technology vendors. As evidence supporting this assertion, I point you to the computer processor vendors. These vendors are falling over one another to improve processor technology, and their press releases are peppered with terms like "unparalleled throughput" and "breakthrough performance."

This behavior is not limited to a few vendors. In the x86 market — the largest and most vocal market — Intel and AMD are battling to shove more cores into their processors — and to do so while lowering power and cooling requirements. In the RISC market IBM is fiddling to find the perfect blend of performance and watts in its Power processors while Sun Microsystems is tinkering with thread counts in its SPARC line of processors. The list goes on — every vendor involved with processor technology is cooking up new recipes for success.

I hate to sound skeptical, but given all of the time, money, and energy that vendors are pouring into processor improvements, we should have discovered something that could accurately be labeled "breakthrough" or "unparalleled."

Yet I haven't heard of any developments that would warrant such illustrious adjectives. In fact, I submit the only significant lessons thus far:

• Processors that operate at faster frequencies provide better performance than those that operate at slower frequencies. We've all been conditioned on this subject — our responses to GHz ratings on processors would make Pavlov proud.
• Processors that operate at faster frequencies have greater power and cooling requirements than those that operate at slower frequencies. Another way to look at this lesson is that the less power and cooling a processor requires, the less performance it offers.
• Adding more processing cores to a processor chip increases overall performance more than adding threads to individual processing cores. (A core is a complete computational unit, and a thread is an execution path within a computational unit).

Now I mean no disrespect to the engineers who are feverishly working on improving processor technology, but aren't these lessons pretty obvious? Couldn't we have figured all of this out by looking at the best computing engine on the planet — the human brain?

Let's contrast the lessons outlined above with what we know about the human brain:

• Frequency — Those of you who consume caffeine know full well that the thought process improves when a stimulant is introduced to the brain. The effect of the stimulant is to speed up thinking, resulting in better overall performance. I could also argue that this explains why "over clocking" (running processors at frequencies higher than recommended) is so dangerous. If you are a coffee drinker, you likely know that there is a line beyond which increased frequency is a bad thing that can lead to computational collapse.
• Cooling — If you've ever had the privilege (or punishment) of living through winter in one of the cold regions of this planet, you know that the colder you get, the slower you think (and move for that matter). Therefore, it should come as no surprise that processors that run at colder temperatures do not offer the same performance as processors that run hotter. I'd rather have my thinking capacity on a warm summer day than a cool winter day. (Here again there are extremes — humans, like processors, can get so hot that they break down and stop working.)
• Threads/cores — Of all the lessons, this one should have been the most obvious. We know that juggling multiple trains of thoughts (the human equivalent of threading) is far less efficient than having some friends help us out with a problem (the human equivalent of multiple cores). The old adage "two heads are better than one" comes to mind; no matter how many ideas we can hold in our head, we can only deal with them one at a time... just like processors.

Perhaps the correlation between human brains and computer processors isn't too surprising when you consider that processors are designed by humans (yes, engineers are human). As the saying goes, "the apple doesn't fall far from the tree." I only hope that my computers don't begin to exhibit the same symptoms of aging that I do — memory lapses are not endearing when they happen inside my hardware.

I believe that our industry would be better served if we limited our use of superlatives to real scientific innovations. Hey — what a breakthrough idea that is; it's an unparalleled notion in our industry!