Why Are PCs and Gadgets So Slow? (Who's Breaking Moore's Law?)

The computer industry needs to stop padding products with ever longer lists of features and "improvements," and start focusing on raw performance.

WEBINAR: On-demand webcast

Next-Generation Applications Require the Power and Performance of Next-Generation Workstations REGISTER >

Posted December 3, 2008

Mike Elgan

Mike Elgan

It's an article of faith that PCs and consumer electronics always get faster and cheaper at a mind-blowing rate.

But that's not really true, is it?

Sure, they get a little faster and a little cheaper, but nothing like the performance and price gains the pundits and vendors promise.

The Gold Standard for predicting performance improvements is the venerable Moore's Law, which says transistor density per dollar doubles every 18 to 24 months. That means, theoretically, that computer chips that cost a given amount of money should double in speed every 1.5 to 2 years.

Intel can demonstrate that its processors have roughly kept up with Moore's Law over the years. So can the many other chip vendors that contribute electronics to your PCs and gadgets.

According to Moore's Law, your PC should be 32 times faster than the PC you used in 1998. And, assuming you buy a new cell phone every two years, each one should be twice as fast as the previous one.

Something is horribly wrong here.

My feeling is that my PC feels a little faster than the one I used in 1998 (as measured by the amount of time I spend waiting for it), but not, say, twice as fast -- and definitely not 32 times as fast, as Mr. Moore might suggest. And my new BlackBerry Pearl feels about 10% -- not 100% -- faster than my old one.

You could argue that everything is better now, that Moore's Law has been applied to a qualitative improvement, rather than directly applicable performance improvements. Software and hardware has evolved, and nobody wants to go back.

After all, what's stopping us from dusting off those old Windows 98 disks and installing them on a 2008 Windows Vista PC? Would that give us 32 times the performance? The answer is: no, not even close. Old operating systems weren't written to take advantage of new microprocessing architectures.

The ugly reality is that there is simply no way for users to realize anything close to the actual performance improvements both promised and achieved by Moore's Law inside chips.

One major reason is that performance improvements at the chip level are squandered by everyone from OS makers to driver writers to hardware makers and component manufacturers. They see Moore's Law as an opportunity to bulk up on new features that motivate buyers to dump old versions and upgrade to new ones.

Performance gains make up for sloppy software development. Software makers see Moore's Law as a benefit for them, not you, so they speed products to market by banging out bloated spaghetti code. The art of concise software writing is dying. Everything is in beta, under construction and scheduled for optimization for a future ship date that never comes.

By the time finished products get to the consumer, the actual user experience is the same as it always was: too slow.

Ten years ago, Microsoft promised a new world of "instant on" PCs and laptops. Where are they?

Sure, Lenovo sells a smattering of "instant on" netbooks. But most of the chatter about "instant on" is by vendors like HP and Dell who are working on, about to ship and on the brink of unleashing a new age of "instant on" laptops and even desktops. But that's exactly where we were in 1998. Back then, "instant on" was available but rare, and a year or two from becoming totally mainstream.

The industry needs to stop padding products with ever longer lists of features and "improvements," and start focusing on raw performance.

If Moore's Law holds, which it will, the speed of processors will double over the next two years. Which would you rather have: Bloated new features that gobble up all those clock cycles before they get to you? Or would you rather have all your stuff run twice as fast?

Industry: It's time to put the feature bloat and spaghetti code on hold for two years, and get to work optimizing, minimizing and streamlining everything. We want speed. And we want it yesterday.

Tags: Microsoft, chips, Intel, Dell, HP

0 Comments (click to add your comment)
Comment and Contribute


(Maximum characters: 1200). You have characters left.



IT Management Daily
Don't miss an article. Subscribe to our newsletter below.