Quad-Core May Be Overkill

In the old days (and by that I mean 2002 or so), we got better performance out of our PCs by simply upping the clock speed of the CPU, with regular architectural updates to add more features to the chip itself.

It wasn't until dual-core came along that things started to change. Today's approach to boosting computer performance is to divide and conquer: Provide two or more "cores," the guts of a microchip, to split a task in two and get the job done without the excess heat and power consumption that comes with skyrocketing CPU speeds.

Of course, there's a bit of a problem with this: Software. Applications haven't really been optimized for multi-core computing, and while some programs excel at the divide-and-conquer approach (depending on how they were written to take advantage of the hyperthreading features of the prior generation of chips), many applications do not. Multi-core chips are a boon to heavily CPU intensive tasks like rendering 3D images or creating video effects, but don't expect Word to spell-check your memo twice as fast just because you have a dual-core PC. Still, the benefits of dual core really do outweigh these drawbacks (particularly in the way they handle multiple applications running at once), and since the price is right I recommend dual-core CPUs for virtually all new computer buyers.

But now there's a new wrinkle, as quad-core arrives next month. If you think your software didn't know what to do with two CPUs just wait until it has four to deal with. Nathan Davis offers some intelligent ranting on why quad-core CPUs are not going to give you much of a performance boost in the real world. Sure, benchmarks look good, but what's the upshot for a day-to-day computer user? Probably nothing except a big fat bill to pay for the latest generation of CPUs.

What's the upshot? By the time software is ready to take advantage of four cores, we'll probably be running eight-core chips. For 99 percent of you, dual core is going to be perfect for the foreseeable future. Give Davis's argument a read for more detail.