Willamette/Rambus

Add Your Comments

.

Even Intel Says Willamette May Burst The Net, but not AMD

It’s never good when you send out slogans rather than hardware to do battle for you.

Intel has basically conceded that the initial Willamettes are not going to be the fastest thing on God’s green earth; that a 1.5Ghz Willamette will probably be slower than any 1.5Ghz competition.

On a technical level, this really shouldn’t come as a shock. Willamette design is essentially the old Alpha “do a little very, very quickly and make it very scalable.”

Willamette wasn’t really designed for 1.5Ghz and .18 micron in mind. It really was designed to work at 2 or 3Ghz, at .13 micron or less.

This is of course Intel being cheap and trying to stretch a design for a long time. That strategy worked well with the PPro, but can it work again?

Waiting for the Lucky .13?

I look at the roadmaps, and look past the hype, and everything points to the middle of next year. PIIIs at .13 micron, the middle of next year. .13 Willamettes, sorry NetBursts, (How low can you go, marketeers? Surprised they didn’t call the ALU Fruit Loops) middle of next year.

Uhhh, this isn’t the middle of next year. What are you going to do over the next nine months?

Apparently, bank on the stupidity and inertia of the buying public. Not as quick as your competitor? Doesn’t matter. As long as you can toss up a big Mhz number and come up with a new buzzword, all is well!

A year ago, I would have been pretty sure this would work. It still might. But I’ve spoken a number of times about the slow erosion of Intel’s name, and it’s happening. The official media is getting more and more skeptical about Intel’s claims, even those considered in Intel’s pocket. They’ve been handed a bill of goods a bit too often, too many
bad motherboards, too many vaporware announcements, too much sizzle and too little beef. What do you think another nine months, especially when faced with real competition, is going to do?

AMD is going to have real, available, cheap 1-1.5Ghz systems available over the next nine months. For the first time, AMD’s going to have real production capacity backing it up. What’s Intel going to have? Words.

Nor will they sit on their laurels. AMD is working on .13 micron, too. By the time Intel gets around to their wunderchild products six months too late; Sledge/Claw/Whatever/Hammer should be close to ready.

This fall/winter was the last opportunity for Intel to kick the AMD child back into the shadows. It isn’t going to happen. This is now a real two-horse race, and the favorite seems to have a wasting disease.

One Little Piece of Good News for Intel

RDRAM prices have been dropping. They’re within 100% of SDRAM prices and continuing to drop. Still probably will never be as cheap as DDR, but we are now faced with the spectacle of Intel rather than Rambus being the trailing party, since Intel doesn’t seem to be too interested in selling a lot of NetBursts anytime soon. The question is: what are they going to sell?

Email Ed


Leave a Reply

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