There Ought To Be A Law

Add Your Comments

From an article in PC World:

“A small group of PC owners has quietly filed a class action lawsuit against Intel, Gateway, and Hewlett-Packard alleging the companies misled them into believing the Pentium 4 was a superior processor to Intel’s own Pentium III and AMD’s Athlon.”

See what happens when you don’t read this website. 🙂

Do you know what’s going to happen here? This will go to court. The plaintiffs will go on and on about being misled about MHz. Intel and Company will produce truckloads of documentation showing how the PIV is better than the PIII, and more truckloads saying that they never said that every PIV would be better than a PIII.

In the end (and it’s not inconceivable the defendants could win a round or two), a judge is going to rule that “yes, this is a bit misleading, but there’s no law (or equivalent) against this.”

The last people who should snicker about this are AMDroids. Just wait until a lawyer specializing in class action finds out a 1.6GHz Clawhammer is called a 3000+ (or if you’d prefer, a 2GHz Clawhammer is really a 3400+). If I were AMD, I’d already have a lawyer working on this one.

And in the end, even though it will be the exact opposite argument, some other judge will say the same damn thing the judge judging Intel said.

Another Standard?

While these court cases will fail per se, they will emphasize and publicize the inadequacy of relying on MHz. This might get the attention of Congress. More likely, it will get the attention of the FTC, who can make regulations that for practical purposes have the force of law.

This should be good for years of Internet fighting. That will be what most people will think is the real issue.

It isn’t. Having the PC industry adapt whatever standard, or having FTC make the PC industry do so doesn’t solve the real issue.

The real story is that AMD and Intel processors are becoming less and less comparable overall and because of that, you cannot have a meaningful single measurement.

Both Hammer and the PIV still have the same x86 roots, but comparing the two will be more like comparing French to Italian than treating them like they still both spoke Latin.

Though they’re both x86 compatible, Hammer and PIV are much different x86 processors, far more so than, say Athlon and PIII.

In this corner, you have a processor with the capability of a form of 64-bit processing, a 12-stage pipeline, strong floating point hardware and a built-in memory controller.

In the other corner, you have a processor with a 20-stage pipeline running a GHz faster than the other contender, software-dependent floating point operations and hyperthreading.

How can you expect such different designs with such different strengths and weakness to perform alike enough for a single measurement? How can one shoe fit two much different feet? Sure, you could come up with a shoe that “fits” all, but if you’re the least bit particular, would you want to wear it?

It’s useless to argue for any kind of official multiple standard. Most computer owners are just too lazy and/or stupid to accept anything more complicated than a single number, no matter how useless, misleading or misguiding it actually is for what they want to do.

Harsh? Maybe, but still so.

I see it all the time. I get emails saying, “What’s the real formula for comparing an Athlon to a PIV?” At this point, that’s just ignorant. I tell them there isn’t a magic formula, that the two chips have rather different performance characteristics, and I tell them to look at benchmarks indicative of what they actually do with their computer. When they don’t want to hear it, then ignorance turns into laziness/stupidity.

I try to explain this to Joe Sixpacks face to face, and if I could glaze glass over as fast as I can glaze eyes with a brief non-technical explanation, Martha Stewart would hire me for some Christmas work. The Joes don’t want to hear it, either.

Look at most of the commercial benchmarks. You used to be able to get per application breakdowns. No more, now you get a single number or a couple, with rare exceptions. Why? Even the 5% (if not less) who’ll make the effort to look at a benchmark don’t want too much of a good thing.

I’m sorry, but when I see people who want a fast computer, but don’t want to hear or make the effort to find out what can make a 30% or more performance difference, that’s stupid and/or lazy.

However, you don’t have to be that way. A little extra homework can pay off a lot, and it will pay more and more as the two major CPU companies continue to make increasingly different products.

Ed

Leave a Reply

Your email address will not be published. Required fields are marked *