A Shot Across The Bow

Add Your Comments

Intel’s CPU strategy explained. –Ed

There’s a bunch of reviews of the 2.0GHz PIV. Yawn.

Oh, the PIV does a little better in general? I should hope so when the Athlon spots it 40% more Mhz.

On the other hand, the Athlon is also spotting that PIV about $450 in price. That makes the price you pay for any improvement you get exorbitant.

Is This Important?

It’s like a wake-up call. Getting a wake-up call when you’re already awake isn’t important at all, it’s only valuable when it comes when you’re not.

This particular wake-up call is really about an hour before you actually needed to get up.

But if “2GHz” and “fastest computer” and “Intel” startles you awake, and you’re too groggy or unaware to hear or understand “Northwood” or “.13 micron”, you might run out and do something before you come to your full senses.

The 2Ghz PIV is the descendant of the 300Mhz PII Klameth and the 600Mhz PIII Katmai, (the 1.13Ghz Coppermine would also be in here if it had ever gotten out). The last Neanderthal before the new Cro-Magnons start showing up.

It’s a dopey purchase when better at the same price is coming in two months. It’s even dopier for overclockers, since the Northwoods are bound to have a lot more overclocking headroom.

When You (Might) Need To Open Your Eyes

The real wake-up call will be Northwood. That’s when performance people will need to wake up for a bit and smell the coffee.

The measurements made then will be important, because then and only then will we get an inkling on how good AMD’s competition is likely to be the next year or two.

Just as AMD had an effective lock on raw performance after the PIII ran out of steam, so will Intel with Northwood until we see Thoroughbred come out.

Northwood will also probably address some the current PIVs failings, plus benefit from increased SSE2 optimizations.

From an overclocking standpoint, Northwood will be the next frontier (though at least initially a very lonely one due to its high cost).

Does That Mean I Have To Buy Intel?

Not at all.

Even from a raw performance standpoint, unless Northwood is much better than we’ve been given reason to believe, Palominos probably won’t trail too much. They will have SSE optimizations, and initial reports indicate they also benefit from SSE2 optimizations.

They may look a bit long in the tooth next spring, and if AMD gets delayed much past next April for the .13 micron Thoroughbred, then things will start looking bad for AMD.

From a price/performance standpoint, though, I don’t see how you can make a case for the PIV for at least the rest of the year, and likely for much longer.

The Worst That Could Come Out Of Left Field

As of now, it looks like Intel will have a leisurely crossover from .18 micron to .13 micron. This is primarily due to problems with buying equipment needed to easily make these chips.

While Intel can work around the problem; it’s not an ideal situation. If you hear they can’t keep up with demand when they have relatively few .13 micron lines available, AMD won’t be in deep trouble.

Should that happen, we won’t see .13 micron PIVs start really selling (which used to happen with overclockers when the price dipped below $250; God only knows what it’ll be now after the past few months) until spring 2002. If AMD can get Thoroughbred out shortly after that, that should even the playing field.

However, Intel has a huge financial interest in getting the PIV line over to .13 micron as quickly as possible. Current PIVs chew up a lot of die space. Intel can make a lot more PIVs a lot better and cheaper at .13.

If you see Intel quickly (by that I mean by January/February) shift over to .13 micron, and rapidly phase out .18; that means manufacturing equipment is not a problem, and that’s very bad news for AMD.

Why? That will mean Intel can sell lower-end PIVs for a lot cheaper than they can now. If most PIVs sold over the next six months are .18 micron, AMD probably can keep prices about what they are now and not get hurt too much. It won’t be easy, but that probably not enough to prevent further expansion.

If we see Intel flinging sub-$100 .13 micron PIVs around early next year, though, AMD is going to have to retrench and fight to survive.

Not saying that’s going to happen. No real indications it’s going to happen. I really doubt that’s going to happen. But if I’m Intel, and could do it, that’s the surprise I’d want to keep absolutely secret until I could spring it on AMD.

The Real Fight Is On the Low End

Intel never has and won’t now sell a lot of $600 processors.

The real purpose of the 2GHz processor is mostly for propaganda purposes. It basically says, “We’re Intel, and we’re back on top,” hoping that that will reassure people buying at the lower end that they’re getting something good for their money when they buy that fairly cheap 1.7GHz PIV/SDRAM system.

It’s the “hit ’em high” part of Intel’s “hit ’em high, hit ’em low” strategy. Get the initiative and the momentum while AMD is perceptually vulnerable and can’t pump up the MHz much, and then follow through with the good stuff.

This is not a battle of realities. This is a perceptual war where most of the participants can barely see.

Intel plans on using Mhz to play Oz. Given the audience, don’t bet the house against Oz.

You want to say that’s bogus, fine, but as we’ve said before, it’s not aimed at you.

It’s aimed at the average people who buys the machines, and if they continue chanting the MHz Mantra, it’ll work.

Like it or not, it’s a very clever strategy. When you have lemons, make lemonade. Can’t say I admire them for a strategy that depends on ignorance, but Intel doesn’t need my admiration or yours. It needs to sell CPUs and whack AMD.

Not saying Intel has to win, but it would be foolish to assume the opposite. Don’t be misled by any personal feelings; this is a major threat to AMD, no matter how stupid, dishonest or flat-out evil it seems to you.

Email Ed

Leave a Reply

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