Forrest Norrod Outlines AMD's Plan To Fight Intel In The Data Center

'We're going to make sure that world-class technology from those companies is extremely well supported on the open EPYC ecosystem, and I think generally that wins,' AMD's top data center executive tells CRN in an interview.

ARTICLE TITLE HERE

Forrest Norrod, AMD's top data center executive, knows the chipmaker can't count on getting its new second-generation EPYC processors into every enterprise's most critical IT infrastructure at the outset.

"When we get a part like [second-generation EPYC] 'Rome,' the thing that will slow us down more than anything else at this point is unfamiliarity," Norrod told CRN in a recent interview.

Instead, to take on Intel's dominance in the data center, AMD plans to target workloads like virtualization, high-performance computing and other areas where the company's new processors particularly shine — and then expand from there, the executive explained.

id
unit-1659132512259
type
Sponsored post

[Related: How AMD Plans To Win Over Solution Providers With EPYC 'Rome']

"The fastest adoption is in areas where our advantage is most acute and where the purchase decisions are most influenced by the technical attributes of the product," Norrod said, referring to workloads for high-performance computing, machine intelligence, cloud and virtualization, big data analytics and software-defined infrastructure. "Any sort of massive scale-out deployment," he added.

AMD launched its second-generation EPYC processors, code-named "Rome" last week, saying that the new CPUs, which feature up to 64 cores, provide a 25 to 50 percent lower total cost of ownership over Intel's second-generation Xeon Scalable processors that launched in April.

While the company is targeting specific workloads to get its foot in the door, the new processors have broad applicability, which is integral to AMD's long-term play, according to Norrod, whose title at AMD is senior vice president and general manager of the company's Datacenter and Embedded Solutions Business Group.

"I want to ramp quickly in those segments with that strategy, and then I want to make sure that we're simultaneously investing such that the longer-to-adopt, but more stable, less lumpy business builds up quickly over the next several quarters as well," he said.

Norrod said the workload-driven strategy is already paying off.

"There's a large automotive manufacturer where we got in on the product design side and the modeling side in their HPC cluster, and now we're everywhere, even on [first-generation EPYC] 'Naples' [processors]," he said.

Patrick Moorhead, president and principal analyst at Moor Insights and Strategy, said AMD and its OEM partners have demonstrated advantages in many different workloads — such as Hadoop RT analytics, Java throughput, fluid dynamics and virtualization — but not in every single area.

Meanwhile, he said, Intel's advantages will rest on newer technologies the semiconductor giant has introduced, such as the Deep Learning Boost instructions included in the latest Xeon processors, which is making a demonstrated impact on low-latency inference workloads. That also includes Intel Optane DC persistent memory, which is well suited for in-memory database workloads like SAP Hana, he added.

"Every ODM, OEM, [cloud service provider] and enterprise I have talked to wants more competition in the space to accelerate innovation and lower costs," Moorhead said in a message to CRN. "With that said, none of these customers would adopt AMD if it didn’t have some advantages."

Norrod said part of AMD's data center strategy involves persuading partners and customers that single-socket servers are adequate for many cases in which they would typically use dual-socket servers. That's because AMD has introduced several new single-socket processors in its second-generation EPYC lineup that the company said can outperform Intel's dual-socket Xeon servers across the stack.

"I don't care what Intel dual-socket server you're talking about. I can beat that performance with a single-socket Rome, offering much lower power consumption, much lower capex, better reliability because there's fewer things to fill," he said.

Intel is aware of the increased competition it faces, but the company recently told CRN that its data-centric platform strategy — which includes an expanding portfolio of CPUs, memory, accelerators and other solutions — is best suited for the modern data center, highlighting several customer and ecosystem wins as evidence.

"It’s natural market behavior to have customers look at their options," an Intel spokesperson said in a statement to CRN. "We are confident in our strategy to deliver a data centric portfolio that serves our customers’ diverse needs, from the core of the data center out to the edge of the network."

Norrod said he knows Intel's argument well, but he questions whether Intel is "world class" in every technology the rival brings to market, pointing to the company's recent admission that it will no longer develop new Omni-Path Architecture products as an example.

"I would suggest that they have not demonstrated the ability to be world class in all of those elements," he said. "And so if you're bundling them together, you're asking the customer to accept the less-than-world-class solution in one or more attributes."

While Intel has a much larger portfolio of products, Norrod said, AMD's strength is in its “open ecosystem” of hardware vendors, which includes Mellanox, Micron, Samsung and Broadcom.

"We're going to make sure that world-class technology from those companies is extremely well supported on the open EPYC ecosystem, and I think generally that wins," he said.

Dominic Daninger, vice president of engineering at Nor-Tech, a Burnsville, Minn.-based HPC system builder who works with both Intel and AMD, said the release of the new EPYC processors is timed well with several of his customers who are planning expansions to their HPC clusters.

"A lot of customers are in expansion, which is an opportunity for AMD," he said.

Daninger said Intel might close the performance gap with its 10-nanometer "Ice Lake" server processors, which the company has teased for 2020, but the company will face challenges in keeping up with AMD's latest advances.

"I don’t see foresee Intel pulling ahead within one generation," he said.

But for all the noise AMD is making about its new advantages over Intel, Norrod said he will never let that make him or anyone else in the company complacent.

"Intel is a great company. They're a formidable competitor. I do not take them lightly at all, and we should never take them lightly," he said. "If my team shows the slightest hint of getting complacent, I quickly disabuse him of that notion."