From the end of this article.
“AMD plans to release about 100,000 K8 processors in the first quarter of 2003, rising to 300,000 units in the second quarter and five million units in the third.”
That one sentence says a ton.
What does it say? It says . . .
1) Clawhammer will really debut in early summer. That’s when it will become commonplace. There’s no way AMD is going to be selling millions of Opterons; the vast majority of these processors have to be Athlon 64s.
2) It will become reasonably priced in the early summer. There’s no way in hell AMD is going to sell five million processors at around $500. At least some of them will have to be reasonably priced. If you respect your wallet, wait.
3) Expect to see multiple bins Expect them to go downward from 3400+ (though see below), and expect them to cover Barton/TBred territory. Before you start salivating, though, the first pioneers will really be taking a big chance; from everything we’ve seen so far, there’s a good chance a lot of these processors really can’t run much faster than their rating.
4) Expect TBred/Barton to be deemphasized in the summer. Five million CPUs is a pretty big chunk of Dresden’s capacity. Not like they’ll stop making Bartons/TBreds; they’ll eventually become the Durons of the AMD family.
Comment: This is hard to believe.
When a new CPU generation is introduced, the period of time it takes from introduction to crossover (i.e, the new product sells more than 50% of the CPUs in its class) usually takes around a year.
What these numbers suggest is that AMD is planning to do this in three months, both in manufacturing and sales.
So in three months, they’ll shift a majority of production from Athlon to Clawhammer. Well, maybe.
In the same three months, the percentage of AMD buyers who buy Clawhammer is supposed to go from around 5% to over 50%. Just like that.
Oh really? Does this sound at all realistic to you?
The last time AMD made a comparable shift, it was from the K6 series to Athlons. There sure wasn’t any sudden product shift like that back then. Took a lot longer than a quarter for Athlons to reach crossover.
I just don’t think people are going to be straining at the bit waiting to buy Clawhammers when older Athlon system will deliver close to same performance.
The only way I see this actually working is if the flagship CPU come the summer is not a 3400+, but the 4000+. Based on what we know about Hammer speeds so far, that would be in the hope and prayer category of possibilities.
I don’t know. . . .