One GHz and Counting

The big news last month was AMD’s

demon-stration of a 1 GHz CPU. Intel soon responded in kind with 1.5 GHz, upping the bar in the chips wars a notch further. And if you were to believe them, they’re doing it all because they have only our–users’–benefit in mind.

Oh! Really?



I’m writing this piece on a four-year old P/100 based notebook, one that I regularly lug all around the country. My other machines are a 133 MHz Mac and a P/200 MMX. Why am I not using faster machines? Surely PCQ can afford a few fast machines. Yes, we can. And we do have faster machines. Lots of them in fact. We use them for software and hardware testing, development work, and design and graphics work, all of which warrant the use of higher processing power.

My problem is that a 500 MHz PIII doesn’t make my typing any faster. A 700 MHz K7 doesn’t retrieve my e-mail any faster. The latest, hottest G4 PowerPC chip doesn’t make my Internet experience any different. Nor do they make any difference to the spreadsheets or databases I use–believe me, I have checked it out. The only noticeable difference was that these applications load a bit faster. If P200’s were available today, I’d happily buy them for our normal productivity needs–e-mail, word processing,

spreadsheets, databases…the lot. The problem is they aren’t available.

Am I against faster machines?



No. The point I’m trying to make is different. And at a time when minimum speeds touch 500 MHz and more, the 30 MHz quarter improvement that vendors tout every quarter works out to a paltry six percent or less over what you’ve already bought. We have long since crossed the stage where improvement in CPU speeds will improve our basic productivity. We have crossed the stage where a better video card can seemingly improve speed of data access from a database.

In such a scenario, would it not make sense to stop these quarterly releases and instead go in for major speed jumps, say with new chips coming out once a year but with significant performance

improvements?

Great as it may sound, I know that this won’t happen. Simply, because that’s not the way the economics of the market works. So, if I want to do nothing more than word processing and e-mail on my PC, which incidentally is true of the bulk of PCs being bought today, particularly by corporates–I’d still have no choice, but to buy a gigahertz machine.

Is there a way, out?



Yes, and No, Yes because there is a set of products that seem to be the answer–thin clients. No, because thin clients today cost pretty much the same, if not more than what a set of individual PCs would. And conventional logic would say, why not go in for the more powerful PC instead? 

Thin clients as a concept has much going for it. There are also some compelling technologies at work. But I haven’t really seen them at work outside research institutes and exhibition registration desks. A quick search at AltaVista threw up many manufacturers of thin clients. There are hardware thin clients. There are software thin clients. There are add-on cards that turn old PCs to thin clients. But none of them seem to offer a compelling product at a compelling price. At least, not yet. And that’s really sad.