AMD seems to have picked itself up off the canvas in its bout with Intel with the launch of the new Ryzen chip, but that is still a far cry from becoming a dominant player in enterprise circles, or indeed knowing whether any x86 architecture will provide optimal support enterprise workloads going forward.
There is more to the life of a processor company than the enterprise, of course, and the initial Ryzen releases are geared more toward gaming systems and PCs than servers. The enterprise will be addressed later this year with the Naples version of the Zen architecture on which the Ryzen is based. But with the high-end, 8-core 7 1800X model coming in at half the price of a similarly configured Core i7, says MarketWatch’s Therese Poletti, we could be on the cusp of an old-fashioned processor war that could lead to significant price drops for a wide range of computing and electronic devices.
The problem wouldn’t be so serious for Intel if AMD had simply stuck to its usual script of offering low-cost, low-performance processors to the bottom of the market, says ZDnet’s Adrian Kingsley-Hughes. But the fact is, these devices outperform their Intel counterparts at a fraction of the price. Under the Cinebench test, the 1800X tops the i7-6900K by 9 percent, while the even lower-priced 1700X edges it by 4 percent. Benchmarks can be tweaked, of course, but the mere fact that the Ryzen comes even close to an i7 at a 50 percent price premium should be enough for OEMs and ODMs to take notice.
AMD has done more than produce a powerful new chip, says ITBE’s Rob Enderle. Under CEO Lisa Su, the company has addressed the reliability and market execution problems that caused many buyers to pull back their support over the past few years. As well, the market for CPUs is rapidly shifting from traditional enterprises to cloud providers, which are much more amenable to experimenting with new hardware architectures due to their greater reliance on abstract infrastructure. By the time Naples hits the channel, AMD should have any number of high-profile deployments showcasing the chip’s ability to handle production-level workloads.
That is, of course, unless something better comes along. One candidate, according to anandtech.com’s Johan De Gelas, is IBM’s Power architecture, which is now available in a 1U form factor from Taiwan’s Tyan Computer Corp. True, the Power8 configuration that Tyan uses is more power-hungry than a typical CPU, but Tyan configures its system with a 750-watt power supply, rather than the typical 1000-watt. And when it comes to the cloud, overall performance in critical areas like HPC, analytics and artificial intelligence usually trumps performance-per-watt. This is part of the reason why Chinese internet behemoth Tencent has opted to deploy OpenPower in its data infrastructure, along with more efficient CPU configurations. (Disclosure: I provide content services to IBM.)
Data requirements are so diverse these days that there is every reason to expect multiple architectures to inhabit a wide range of infrastructure. But with the steady commoditization of hardware taking place, price pressure will be intense as providers seek to deliver the highest performance at the lowest cost.
For chip designers of all stripes, then, the justification for high-end models will have to be pretty strong if they hope to maintain share of a market that is becoming increasingly less interested in the inner workings of the hardware it uses.
Arthur Cole writes about infrastructure for IT Business Edge. Cole has been covering the high-tech media and computing industries for more than 20 years, having served as editor of TV Technology, Video Technology News, Internet News and Multimedia Weekly. His contributions have appeared in Communications Today and Enterprise Networking Planet and as web content for numerous high-tech clients like TwinStrata and Carpathia. Follow Art on Twitter @acole602.