AMD chip strategy shakes Intel to the core

WE DON'T KNOW yet how far advanced Intel's plans for its dual core processor are, but rumours were reaching the INQUIRER late last year that the firm was re-jigging its entire roadmap after realising it had hit the glass ceiling on processor performance.

So the confirmation yesterday that AMD has dual core desktop chips on its roadmap must surely be concentrating Intel's mind and focusing its engineers' attention on how it will combat the threat from its much smaller competitor.

So far, Intel has been pretty vague about what it will do on the dual core front in 2005. It's still unclear whether it will throw the Prescott design away, rip up the roadmap, and re-build its architecture using technology grown at its Israel R&D centers.

It's performing a juggling act in the run up to the launch of its LGA 775 processors, which may not be available in as big quantities as you'd expect after the launch on June 21st.

What's clear is that both AMD and Intel are playing a high stakes poker game in the chip casino. Neither wants to show its hand just yet, but each is raising the stakes in an elaborate game of bluff.

While Intel continues to push its Extreme Edition Pentium 4 as the high end answer to an enthusiast dream, AMD continues to keep the prices of its desktop FX series deliberately high. Neither firm will say how many of these chips have been sold, but that scarcely matters in this high roller game.

Intel, as we revealed earlier this week, has begun to sprout a range of workstation/server chips using iAMD64 but without much of a song and dance. It still seems to be firmly focused on the IA32 architecture, at least in the short term.

Intel has always maintained that the 64-bit market will not really take off until Microsoft launches its 64-bit version of Windows XP. When the software firm deigns to do so, we will quite probably see a swift change of direction from Intel.

Despite Linux alternatives being available for the 64-bit desktop processor market, it is an undeniable fact that these chips will only start to gain traction when the MS OS becomes reality and drivers start becoming widespread.

This standoff may benefit Intel in the long run - its competitor may be right that 64-bit PCs are the shape of things to come, but without an OS and without compelling applications, we may all end up having very powerful tin that does not really do much more than our existing 32-bit boxes.

Intel president Paul Otellini said in a call to analysts earlier this year that he did not expect iAMD64 to take off until such applications were widespread. He suggested that 64-bit Office might be a compelling app eventually, but we think he was being deeply ironic here. The pixels on your word processor may shuffle about a bit faster, but the magic of huge amounts of memory and 64-bit power will not do anything for your typing speed. A course in touch typing is likely to be more effective and cheaper than investing in the superset of hardware right now.

Jerry Sanders, founder of AMD, once said that with Intel and Microsoft, his firm was the holy trinity. By that he means that X86 architecture and Windows software are the things that really shift boxes.

But it does seem to us here at the INQ that Microsoft may well be using the whole 64-bit market as a way of quietly diminishing the power of Intel.

The game is far from over yet. Intel has massively more resources than AMD and when Otellini displaces Craig Barrett as CEO of the chip giant, we suspect that AMD will find itself once more facing an aggressive, re-invigorated and determined Chipzilla.

It would be a real mistake for AMD to think it has a real edge in a battle which may only actually be a border skirmish so far.