Skip navigation
Intel Flashes 10nm Nexg-Gen Xeon Phi at Year's Big Supercomputing Show
Upcoming Knights Landing Xeon Phi (pictured) will be first to use silicon photonics, but the generation after that (Knights Hill) will use Omni-Path, which the company says is faster than InfiniBand (Photo: Intel)

Intel Flashes 10nm Nexg-Gen Xeon Phi at Year's Big Supercomputing Show

Supercomputer chips will use Omni-Path interconnect technology, Intel claims is faster than InfiniBand

At this week's SC14 in New Orleans, the big supercomputing show of the year, Intel selected a handful of details to disclose about the future of its Xeon Phi line, code-named Knights Hill, as well as new architectural details around its Omni-Path fabric interconnect technology. As the Xeon Phi series matures and more systems are deployed with it, Intel is assuring all stakeholders that there is a future for the chip that is becoming a standard high performance computing building block for some of the world's fastest supercomputers.

Following the upcoming 14 nanometer Knights Landing product, Intel said its third-generation Knights Hill Xeon Phi family will be built using 10 nm process technology and integrate second-generation Omni-Path Fabric technology.

The first commercial systems based on Knights Landing are expected to ship next year and will incorporate Intel’s silicon photonics technology. This will also be the first stand-alone Phi chip in the product line which has consisted of co-processors, used to offload calculations from the main CPU.

While sticking to its iterative release process for Xeon Phi, Intel has shown flexibility with custom Xeon orders, making special chips for new Amazon C4 instances and for Oracle database machines.

According to the November 2014 Top500 list Intel-based systems account for 86 percent of all supercomputers and 97 percent of all new additions.

Intel has re-branded its OmniScale interconnect technology to Omni-Path and said the new architecture was expected to offer 100 Gbps line speed and up to 56 percent lower switch fabric latency in medium-to-large clusters than InfiniBand alternatives. The Omni-Path architecture is based on a 48 port switch with 100Gbps line speed.

To further enable Omni-Path Intel launched a Fabric Builders program to help form an ecosystem to act as a catalyst for working together on Omni-Path solutions.

Global geosciences company DownUnder recently purchased a customized SGI Rackable HPC solution with Intel Xeon processors and 3,800 Xeon Phi co-processors. DownUnder managing director Dr. Matt Lamont said the "combination of Intel Xeon Phi co-processors with our proprietary software allows us to provide our customers with one of the most powerful geo-processing production systems to date.

"Our Intel Xeon Phi powered solutions enable interactive processing and imaging from each of our geophysicists' individual computers. A testing regime that once took weeks can now be achieved in days."

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.