StorageAMD Intros First ARM-Based Opteron Server Chip

AMD Intros First ARM-Based Opteron Server Chip

ServerWatch content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

Advanced Micro Devices continues to lay the groundwork for its ARM-based server processor plans, unveiling its upcoming eight-core Opteron A1100 Series “Seattle” chip and a development kit complete with an open-source software stack.

Also at the Open Compute Summit event in San Jose, Calif., Jan. 28, AMD officials announced the company is contributing a microserver design that uses the Seattle system-on-a-chip (SoC) that will fit into the Open Compute Project’s (OCP) motherboard design, called “Group Hug.”

The company’s strategy will continue later this quarter when Seattle begins sampling, as will a development platform designed to make it easier for programmers to design software for AMD’s ARM-based platform, according to Andrew Feldman, corporate vice president and general manager of AMD’s Server Business Unit.

AMD officials announced in 2012 their intention to produce 64-bit ARM-based server chips, giving the company a portfolio that also will include its x86 processors. In a press conference at the show, Feldman said the growing numbers of people getting onto the Internet, the increasing numbers of connected mobile devices and the rise of hyperscale data centers are driving a demand for cost-effective and energy-efficient alternatives to the x86-based chips built by Intel and AMD.

“These new workloads mean that the same-old way of doing things just doesn’t work,” he said.

Feldman has been among the most vocal supporters of the push to bring ARM into the data center, and repeated his belief that within a few years, ARM-based systems will grab a significant share of the global server market—25 percent by 2019—and that AMD will be the world’s largest maker of ARM-based server chips.

That will start with the A110, a four- or eight-core SoC based on ARM’s Cortex-A57 design that is built on a 28-nanometer manufacturing process. They’ll run as fast as 2GHz and offer up to 4MB of Level 2 and 8MB of Level 3 shared cache.

AMD’s Opteron A-Series development kit, packaged in a Micro-ATX form factor, includes not only the A1100 Series chip but also a Fedora Linux environment, Linux GNU tools, Apache Web server, a MySQL database engine and the PHP scripting language. Java 7 and Java 8 will offer developers the capability for working in a 64-bit ARM environment.

In addition, there are four DIMM slots for up to 128GB of DDR3 memory, PCIe connectors, SATA ports and 10 Gigabit Ethernet support.

ARM, AMD and other ARM partners—such as AppliedMicro and Marvell Technology—see an opportunity with the growth of hyperscale data centers to leverage the performance and power efficiencies in an ARM architecture that grew up in mobile and embedded devices for low-power servers. Some top-tier OEMs—particularly Hewlett-Packard and Dell—have signed onto the vision with plans for ARM-based microservers.

HP has announced that it will ship ARM-based microservers in 2014, though the first of HP’s Moonshot systems are running on Intel Atom SoCs.

This year is shaping up to be a crucial one in this push. ARM is scheduled to come out with its ARMv8 design, its first architecture to offer 64-bit capabilities, as well as other server-level features like more memory and greater support for virtualization. Feldman and other proponents expect systems to hit the market later this year, with momentum gaining speed in 2015.

However, the ARM camp took a hit last month when one of the earliest ARM-based server chip makers, Calxeda, suddenly shut its doors after essentially running out of money, even after entering into a high-profile relationship with HP.

Intel also is pushing its x86-based products in the low-power server space. The chip maker in September 2013 launched the Atom C2000 “Avoton” family of SoCs—the second generation of Atom-based server chips—that is based on the company’s new “Silvermont” architecture that Intel officials say meets or exceeds the performance and power efficiency of ARM-designed offerings.

Intel officials also argue that the company’s products offer programmers a familiar x86 development platform that lessens the need for reworking software to run on a new architecture.

Feldman and Ian Drew, ARM’s chief marketing officer, both pointed to the growing open-source ecosystem growing up around the ARM architecture, and noted the work of the Linaro Enterprise Group in encouraging programmers to write for the platform.

Get the Free Newsletter!

Subscribe to Daily Tech Insider for top news, trends & analysis

Latest Posts

Related Stories