OPINION: 64-Bit CPU Showdown

Despite what Intel must sure wish, AMD isn't likely to go away, nor does the challenger seem inclined to fall back and let Intel reassert its one-time dominance. Most analysts and benchmark freaks agree that in the 32-bit CPU market, AMD's Athlon and Duron processors mount a credible, if not superior option to Intel's Pentium and Celeron lines. However, many don't even give the upstart more than a slight chance to compete with Intel's upcoming 64-bit Itanium processor.

If you've been following the progress of the Itanium, you've surely heard that it promises Intel's first break from the basic 8088/8086 Intel architechure(IA). Of course, a new architechure means new native software. The Itanium will reportedly run standard IA code in a 32-bit emulation mode that will be hardwired into the CPU. To reap the benefits of 64-bit processing, applications and OS code will need to be compiled specifically for the Itanium.

This is a risk for Intel, but one that probably is inevitable, at least eventually. Motorola made a similar switch sucessfully from the 68xxx series to the PowerPC processor a few years back. Technological advancements make it necessary and even desirable to occassionally make a clean break with the past. However, the timing of these breaks is critical.

Motorola had no competition for its established Apple Macintosh market, ensuring that no upstart would come in and produce a "super 68xxx" processor to compete with the paradigm-breaking PowerPC. Intel doesn't have it so easy. AMD, which doesn't promote or even discuss future CPUs much, has ackowledged that it is developing a 64-bit CPU of its own. In an ironic twist, AMD's chip is reportedly Intel-architechure derived--meaning that by today's standards, AMD will be Intel-compatible, while Intel won't be.

This promises to be a huge battle with plenty more ironic twists. If Itanium launches with exceptional 64-bit performance and adequate "legacy" performance, AMD will likely counter with adequate 64-bit performance and exceptional legacy results. Developers, long an Intel strength, will be asked to make a choice--write 64-bit applications that only run on the Itanium or create perhaps slightly less efficient 64-bit applications that can conceivably also run on the huge base of 32-bit IA processors, including Intel's own upcoming Pentium 4 line.

Early reaction to this potential battle has been relatively sparse, and in some instances, quite baffling. Some pundits have already declared the Itanium the winner, citing its "overwhelming" lead in attacting native development efforts. Ovbiously, they've missed the point. Native IA-64 applications wouldn't really need to be in place for the potential AMD chip because it would run virtually all the huge library of current applications. If consumers buy into the AMD CPU, native applications would surely follow. The other point these pundits miss is the relative ease of writing for a derivitive CPU as opposed to a whole new CPU. Programmers with years and even decades of experience with the Intel architechure are likely to be attracted to an option that leverages that experience.

Its way too early to call a winner. Intel is a powerful company with incredible resources, both technological and organizational. AMD on the other hand has the advantage of being able to counter-punch. Everyone expects Intel to lead, whenever AMD pulls ahead in the CPU race, no matter how slight, Intel takes a PR hit. However, AMD takes no such damage from Intel's jumps into the lead.

64-bit CPUs are probably further in most people's desktop future than even their eventual release. Intel is propogating the notion that the Itanium is primarily designed for server, massive database, and transaction handling installations and that the company's Pentium line of 32-bit CPUs will continue to evolve for the mid-range workstation and consumer markets.

If you must bet on one or the other winning, go ahead, but just make sure that you don't take steep odds. The challenger may not be getting all the advance notice, but AMD is quietly positioning itself to take a real shot at taking over Intel's role as the leader in mass-market CPU technology.