Intel has apparently had some awkward moments during its forum.
During one of the presentations, one of the presenters apparently asked the audience how many Willies they had, and got the willies when hardly anyone raised their hands.
They had an Itanium Forum, and somebody asked why they should buy an Itanium now rather than a McKinley later. The normally voluble Intel panel got quiet.
Not that McKinley is going to be fighting for sales against Durons. The prototype dual-processor McKinley demonstrated required a hefty 600-watt power supply to run. It has a CPU fan spinning at 70cfm. It has eight DIMMS slots. All of this reminds you of one thing:
This isn’t meant for you.
Intel didn’t build Itanium or McKinley as a further development of their mainstream processor line. They built it to compete against people like Sun in servers.
But has it been worth it? How much money does Sun make compared to Intel? They’re hardly hurting, but the gross margins aren’t better than Intel’s.
More importantly, how much has the drain of design and technical resources cost Intel in its bread-and-butter business? If Intel were completely honest, Coppermine should have been called, “We’re too busy building Itanium.”
If Intel had focussed its technical resources on the mainstream products that bring home the bacon, AMD would be considered a misspelling of “and” rather than what it is today.
A focused Intel could have crushed AMD numerous times during its resurgence, but they’ve been doing everything else but that. They think they’re invulnerable.
It’s so reminiscent of American auto makers in the late sixties, early seventies, “We can blow this cheap Japanese imports away any time we feel like it.” Did they?
Let’s jump ahead a year. It’s 2002. AMD gets Clawhammer and Sledgehammer out. They’re priced right enough for everybody to buy. They’ll run current 32-bit applications really fast, and can handle 64-bit ones, too. Probably not quite as well as Intel’s but the average person won’t know that, or care if they do.
What does Intel have to compete against it over most of the product range? Nothing. Oh, Willy2. Let’s assume Intel blunders one more time and essentially just shrinks the current Willy1. This is a chip that can’t even beat current Athlons, not even in FP with SSE2 support.
How do you think it’s going to look against a Clawhammer with SSE/SSE2 support plus superior fp hardware?
Oh, someday Intel will migrate their 64-bit solution down to the masses. 2003, 2004. That gives AMD at least a year’s headstart.
AMD will be well on its way to building a third fab, and if Clawhammer really takes off, I’m sure they’ll be able to get some other fabbers to help them out.
And when the Intel solution finally arrives, it will probably cost a lot more, handle about 12 applications wonderfully, and botch the other 120,000 out there.
Am I the only person who thinks this isn’t too bright?
And please, I don’t want to hear from, “I work in the IT industry and we want Intel, case closed.” Maybe in your little neck of the woods, but this is a forest, and AMD may turn into Robin Hood.
Outside of your little world, it’s the Joe SixPacks and the corporate beancounters who rule. They buy a lot more than you, and they’ll decide this one for the mainstream.
Not saying Intel will lose, just saying they can, and you’re deluding yourselves if you think otherwise.
Tags: Systems & Components