AMD window opens as Intel Tejas closes

CHIP FIRM AMD probably has a clear lead of over 12 months and as long as 18 months to make hay while the sun shines on the performance stakes, following Intel's decision to can Tejas and Jayhawk microprocessors last week.

It could have longer. But that will depend on how Intel re-markets its dual core chips and how quickly it can move to a 65 nanometre process for the Prescott Pentium 4 platform. It's clear that Intel already has started messing with the next process shrink and the firm was at pains to tell us last week that Grantsdale and Alderwood, its corporate stable platforms, are likely to take advantage of such size reductions on future microprocessors.

While we've been reporting since the beginning of the year that the Pentium M core is the shape of things to come, Intel still formally refuses to comment on whether its dual core "home/office" introduction "brought forward to 2005" is a Pentium M.

But this poses another question - one which we asked Intel last week and which it couldn't satisfactorily answer. Generally speaking, when big announcements like this are made, Intel prepares a Q&A with all the likely questions on it but one we asked appeared to catch the firm unawares.

If you're a home and office user, is it better to buy a Pentium 4 Prescott or a dual core Whateveritis? In other words, which is the better performer? That's given that we're going to have to rely on Intel's numbering scheme to figure out what's what, and those numbers aren't really a performance rating, more an overall guide to general functions.

Are we going to get a dual Pentium M a dual number such as 732-732? No one knows yet, and that, we suspect, includes Intel itself.

Maybe Intel will make life easier for the motherboard makers and system integrators by bunging the dual core son-of-Pentium M into an LGA775 package.

Where does all of this leave Advanced Micro Devices (AMD)? Possibly in a strong position. It appears to be able to migrate its microprocessors to 90 nanometres, the roadmap seems to be more or less intact, and it's promising both the K9 and the 65 nanometre shrink too.

Until Conroe appears, it's probable that AMD could well have a window of opportunity provided it doesn't screw up on production, its partner relationships, and a heap of other imponderables too. Let us never underestimate AMD's ability to shoot itself in the foot. µ

See Also
Brookwood pips Glaskowsky in Tejas quote race
Intel confirms Tejas, Jayhawk dead
Intel may have canned Tejas project
Intel's up to date roadmaps
Intel prepares to can Pentium 4