Intel and AMD are locked in a 64-bit processor shipment competition that resembles a race between two broken-legged fillies.
Horse number one is Intel's Itanium processor, which has stumbled along since its birth. After enjoying a modicum of success with the second generation McKinley chip, Intel saw sales slip in Q1 and any momentum was lost.
More shocking, however, has been the start of AMD's Opteron chip - known here as horse number two. This chip has everything that Itanic lacks. It offers 64 bits but maintains the x86 instruction set as opposed to Itanium's EPIC. The price for an Opteron is far below (cut more today) that of an Itanic. And most of all, the public perception of Opteron is more favorable than the cool reception Intel's beast tends to receive.
All of these factors would seem to lead to a robust opening for Opteron, but instead, we find AMD's CEO Hector Ruiz hiding behind a dodgy bean-counting contest that shows how poorly the chip has done thus far.
For example, The Austin American Statesman reports Ruiz making the following bold claim during AMD's recent earnings announcement:
"We have shipped more 64-bit Opteron processors (since it was introduced in late April) than our competitor shipped 64-bit processors in all of 2002," Ruiz said.
The paper then goes on to state that AMD would not detail specific Opteron figures. No surprise there, but so what? The really bizarre part of the statement is that Ruiz compared Opteron to Itanic at all.
Ruiz has thrust AMD right into the center of a "whose chip is selling the slowest" contest. Horse number one has just careened into the fence, and the owner of horse number two is jumping for joy because his beast might just stay upright on the three good legs it has left.
If AMD really wants to compete with the big boys, it should start comparing sales with 64-bit chips from IBM, Sun and HP. Or better yet, face up to reality and pin Opteron against the Xeon processor, which is a more apt comparison.
These are harsh words for a chip that shows much promise, but they are necessary given AMD's marketing tactics. The company has decided to use vague, unsubstantiated numeric claims to offset the apparent early woes of its bet-the-company processor. Every new technology needs time to gain a place in the market, but AMD should do a little more to face the facts and present a realistic picture to customers.
The fall of leading Opteron server maker Newisys should have sent a signal to the AMD brass to lay off the braggadocio for a bit. Newisys was the company meant to bring IBM, Dell and others on board with Opteron. Newisys hinted to the press that OEM interest abounded, and reporters here and elsewhere swallowed the line whole. The sale of Newisys to Sanmina, however, confirms that revenue came in far slower than hoped.
Even if Opteron is outpacing Itanic, which we doubt, it's doing so with large, supercomputer wins carrying the load. At least Itanium has managed to crawl into the enterprise, albeit one hoof at a time.
In addition, Intel has just released the third generation Madison into the wild. While the chip might not generate incredible sales, it's sure to outpace its predecessors.
Bless AMD for keeping Intel honest as a first-rate competitor. Opteron's success could do great things for the industry by making 64bit computing available to companies of all sizes and perhaps sparking a new wave of innovation.
Let's just hope the company's hype doesn't cut another leg out from under its thoroughbred before the horse has even left the starting gate. ®