Will Opteron's first birthday be its most memorable? [printer-friendly] • The Register

Original URL: http://www.theregister.co.uk/2004/04/23/amd_opteron_birthday/

Will Opteron's first birthday be its most memorable?

Dream on...

By Ashlee Vance in Chicago (ashlee.vance@theregister.co.uk)

Posted in Servers, 23rd April 2004 09:30 GMT

Analysis AMD's Opteron processor has enjoyed more success in one year than most thought possible, but any notion of a guaranteed, glorious future should be dismissed.

It's easy to be impressed by the gains AMD has made in the server market over the past 52 weeks. This time last year, AMD was a vague afterthought in the minds of server customers. Some used Athlon processors in low-end boxes, but that's where AMD's traction stopped.

Opteron changed everything in the enterprise market for AMD. IBM was the first big name to pick up the chip with a server aimed at high performance computing, and then HP and Sun outdid their rival by announcing entire Opteron server lines. The backing of major vendors has helped AMD make a quick move from the niche technical computing server market where new chips often languish into real-life data centers.

This is no small achievement for an unknown quantity - just ask those responsible for hawking Intel's Itanium processor.

AMD even managed to line up several large software makers behind Opteron. CA, IBM, Microsoft, Oracle, Red Hat, SuSE, Sun and smaller guys such as VMware are on board.

And at the bottom line, where things matter most, AMD recently pumped out a quarterly profit (http://www.theregister.co.uk/2004/04/14/amd_q104_results/) for the second time in a row. Opteron sales helped push AMD's average processor price higher, which clearly fattened total revenue.

Overall, AMD shattered industry expectations for Opteron's first year. The fact that the company's product is even discussed as a serious Intel alternative in the server market is a major accomplishment. But a huge question remains as to how much of the server market AMD can hold onto as Intel rolls out its own Extended Memory 64 Technology (EM64T) Xeon processor in the very near future.

The truth is that as awesome as AMD's one year gains are, they are not enough to knock Intel off its pedestal anytime soon.

We have a pool and a pond. The pond is probably best for you.

From the hardware vendors' perspective, Opteron is still a second class chip.

After one year, IBM continues to relegate Opteron to a high performance technical computing box that fills a small void in Big Blue's overall server line. IBM's main priority remains its own Power processors - a line the company hopes to promote in 2004 as a better than ever choice for low-end servers and blade systems. This is Opteron's prime stomping ground. In addition, IBM must also back a vast Xeon franchise and nudge Itanic where needed.

HP has its own existing businesses to protect. The company owns the majority of the Xeon server market and always has its huge commitment to Itanium in mind. True, HP has rolled out an impressive Opteron lineup, including the introduction of a four-way box this week (http://www.theregister.co.uk/2004/04/22/hp_ships_fourprocopteron/), but these point products hardly rival the completeness of HP's Intel-based product line. This may change over time, but then again maybe not. HP has been a big backer of AMD on the PC, but Intel-based sales still make up the vast majority of its personal computer business.

Sun's case is a bit less complex. Sun appears more than happy to back AMD over Intel in the x86 market and plans to roll out workstations, a four-way box and an eight-way box to complement an existing two processor system. Sadly for AMD, Sun's place in the x86 market defines the word niche.

And then there is Dell. Can AMD be considered a major threat in the x86 server segment without Dell's backing? Not really.

Come on, Michael. Intel will never know what we've done.

Dell's importance is two-fold. The most obvious need for Dell support comes from the incredible volume of servers sold by the company. While its mostly a perception issue, having Dell on your side certifies your product's position as a low-cost, practical option. Dell is in some ways the lowest common denominator - a marker that says it makes sense to give this product a try.

But there is another more superficial component to Dell's backing that is key.

On the PC, AMD managed to carve out a large chunk of the market, but its share is still surely a joke at Intel. Large OEMs flirt with Athlon but don't manage to ship loads of AMD-based gear. That's because AMD is not really essential. It's interesting and cheap but not essential. If it were essential, Dell would sell the product.

This is the hump AMD must get over with Opteron. AMD must prove that Opteron is not just competitive with Xeon Extender but superior. If Dell started shipping Opteron boxes, it would signal to the market that customers are asking for Opteron by name over Xeon, making it essential.

To its credit, AMD has a chance to pull off this feat. As HP proved again this week with the new four-processor box, Opteron is outpacing Xeon on numerous benchmarks. And that's just with 32-bit software. Athlon rivaled Pentium in the past too but not with equally impressive results. If AMD can continue to truly outclass Intel, it will win Dell and the market's approval.

The problem for AMD is that Intel has a fair amount of time to catch up. As good as AMD's performance is on 32-bit software, Intel already has that market more or less locked up. Customers may flirt with a few Opteron boxes for 32-bit code, but they're not likely to mess too much with what's working.

On the 64-bit side, AMD has to wait with the rest of us for the software market to mature. The Linux OS vendors are there, but Microsoft is taking its sweet time rolling out a 64-bit version of Windows for AMD64. And that's just the operating systems. Hell, VMware made AMD's Opteron birthday press release, and the ISV doesn't expect to have a complete line of AMD64 software for 18 months (http://www.theregister.com/2004/04/19/vmware_supports_opteronxeon/).

It would be great for the market if we could still be touting AMD's 64-bit lead over Intel 18 months from now. AMD's success can only mean better prices and better products for customers.

For this to happen, however, AMD can make no mistakes. It's a familiar story for the vendor, but the stakes have never been higher. Unlike times past, AMD has beat Intel to the prize, meaning its has everything to lose and everything to gain. ®