Intel and ARM continue to grow the fields in which they compete, from mobile devices to the data center.
When ARM officials several years ago first broached the idea of using the company’s low-power system-on-a-chip (SoC) architecture to challenge Intel’s dominance in data center servers, the focus was on highly dense and highly energy-efficient microservers that would find their ways into the massive environments run by giant Web companies like Google, Facebook and Amazon.
However, at this week’s International Supercomputing Conference (ISC) 2014 in Germany, ARM chip makers Applied Micro and Nvidia move to push the ARM architecture—which is currently found in most smartphones and tablets—into the high-performance computing (HPC) realm. Applied Micro officials said June 23 that the company’s 64-bit X-Gene SoC, which boasts a custom CPU built on the ARM architecture, has the high performance, throughput and energy efficiency needed for supercomputer workloads.
At the same time, Nvidia officials said the company’s GPU accelerators—used in many HPC systems to boost performance while keeping power consumption in check—support 64-bit ARM processors.
For its part, Intel is not taking its dominant position in the supercomputer space for granted. The company on June 23 unveiled new details on the upcoming next generation of its Xeon Phi supercomputer chip, which will offer three times the performance of the current Xeon Phi “Knights Corner” offering and could eventually find its way into smaller servers and possibly even mobile devices.
The upcoming Xeon Phi “Knights Landing” chip, which will begin appearing in systems in the second half of 2015, is an important step in the push toward exascale computing, according to Charles Wuischpard, vice president and general manager of workstation and high-performance computing at Intel.
“The race to exascale by the end of the decade is one of the goals we all have … in the high-performance computing market,” Wuischpard said during a press briefing before the ISC opened.
Intel over the past several years has been aggressive in broadening its reach beyond the core PC and data center server spaces. It has worked to drive up the energy efficiency of its Atom SoC platform to enable it to compete with ARM processors made by the likes of Samsung and Qualcomm in smartphones and tablets, with mixed success. ARM has a dominant position in the mobile device space.
Conversely, ARM faces an uphill climb as it looks to chip away at Intel’s huge share of the server chip market. ARM officials have talked about the need for greater energy efficiency in an era of mobile computing, Web 2.0 workloads and the cloud, and the desire of organizations for an alternative to x86 chips from Intel and Advanced Micro Devices.
With ARM’s partners now able to leverage the chip designer’s 64-bit ARMv8-A design, the push is becoming less theoretical and more practical. However, while most ARM partners are looking to offer SoCs for microservers by building their chips based on ARM’s off-the-shelf CPU design, Applied Micro and others, such as Cavium, have taken their ARM license and built their own CPUs designed to compete with Intel’s mainstream Xeon server chips.
Cavium earlier this month at Computex unveiled its upcoming Thunder X ARM-based server chips. “Most [businesses] are looking for a second vendor to Intel,” Gopal Hegde, vice president and general manager of Cavium’s Server Processor Group, told eWEEK at the time. “We need to offer them a choice.”
Applied Micro for several years has been developing its Gene-X processors. Its X-Gene 1 SoC is now shipping to system makers, while its 28-nanometer X-Gene 2 is sampling with OEMs, according to Guarav Singh, vice president of strategy for X-Gene and product lines.
At the ISC, Applied Micro is demonstrating systems powered by X-Gene, and announced that it will work with E4 Computing Engineering to develop low-power EK003 servers powered by the ARM chip. In addition, a few days before the show, Softiron announced the X-Gene-powered 64-0800 enterprise-class server motherboard.
Intel, ARM Take Competition Into HPC Arena
Meanwhile, Nvidia officials announced that its Tesla K20 GPU accelerators with the vendor’s CUDA 6.5 parallel-processing platform support X-Gene in systems aimed at the HPC space, and that the first GPU-accelerated ARM64 development platforms will be rolling out from system makers like E4 and Cirrascale in July.
Officials with both Applied Micro and Nvidia said 64-bit ARM SoCs are a good fit for the HPC space, where power efficiency is increasingly important and recompiling software for particular architectures is nothing new.
“There is definitely a need for power efficiency, so HPC guys are incentivized” to explore alternatives to x86 chips, Singh told eWEEK. Roy Kim, marketing manager for Nvidia’s Tesla Group, agreed.
“HPC is definitely the most flexible out of all data center customers,” Kim told eWEEK. “They’re used to recompiling [their applications] all the time.”
However, with Intel, ARM and its partners are competing against a highly innovated, well-funded and determined company. Intel has leveraged its Atom platform to contend with ARM’s expected encroachment in the dense, low-power server space. Now the company is making significant strides with its next-generation Xeon Phi chips.
The new Knights Landing chips will include the company’s Silvermont microarchitecture, which currently is found in Atom SoCs and will help drive up the performance of the processors while reducing the power consumption. It also includes new on-package memory being developed with Micron Technology that will offer five times the bandwidth of DDR4 in a third of the space and with five times the power efficiency. It will start off with 16GB in the new memory.
At the same time, Knights Landing will feature a new fabric technology that Intel officials call Omni Scale that also will be integrated with the 14nm generation after Knights Landing. Rajeeb Hazra, vice president of Intel’s Data Center Group and general manager of its Technical Computing Group, would not go into many details about Omni Scale, saying only that it wasn’t Infiniband—which Intel is using today—but is compatible with it. Creating an interconnect fabric has been a focus of Intel for several years, and a driving force behind such acquisitions as technologies from Cray and QLogic, as well as Fulcrum Microsystems.
Intel’s Wuischpard also said that the Omni Scale fabric technology will be open, enabling other vendors to build off of it.
The Intel officials said Knights Landing will have more compute cores than the 60-plus in the current Knights Corner, but would not say exactly how many. However, with all the enhancements and new technologies, they said it will offer more than 3 teraflops of performance in a single package, more than the 1.2TF in the current generation of Xeon Phi.
The key is moving beyond just offering a compute core by integrating everything from the interconnect and power efficiency capabilities to the memory and storage onto a single chips, Hazra said.
“It’s not just the processor performance, but all these other parts as well,” he said.