Skip navigation
Intel Phi Coprocessor Released With Broad Industry Support
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 Phi Coprocessor Released With Broad Industry Support

Intel announced the commercial release of the Xeon Phi Coprocessor, that when combined with Xeon processors promise to drive innovation in high performance computing (HPC).

Intel announced the commercial release of the Xeon Phi Coprocessor, that when combined with Xeon processors promise to drive innovation in high performance computing (HPC). The announcement was supported by many vendors, including SGI, Cirrascale and Appro.

Intel (INTC) officially introduced the Intel Xeon Phi coprocessor at the Supercomputing Conference (SC12) in Salt Lake City Monday. Intel's Xeon Phi was formerly known by the code name Knights Corner, featuring 3-D tri-gate transistors and 22-nanometer silicon. Most important in the new coprocessor is its familiarity in programmability, extending the Intel Xeon processor architecture. Specializing in efficient performance for highly parallel applications the Xeon Phi boasts up to 8GB of memory and 352 Gigabytes per second of peak memory bandwidth.

"Intel Xeon Phi coprocessor represents an achievement in Intel innovation that will help propel us to new heights in research and discovery, and reaffirms our commitment to Exascale-level computing," said Diane Bryant, vice president and general manager of the Datacenter and Connected Systems Group. "The combination of the Intel Xeon processor family and the Intel Xeon Phi coprocessor will change the scope and scale of what highly parallel applications can accomplish, by delivering unprecedented performance, efficiency and programmability. With this technology as a new foundation for HPC, solving real-world challenges from accurately predicting weather patterns 21 days in advance, to developing new cures for diseases will become increasingly possible."

Intel launched two Xeon Phi product families to provide optimal performance and performance-per-watt for highly parallel HPC workloads. The Xeon Phi coprocessor 3100 family will offer more than 1 Teraflop/s performance, up to 6GB memory at 240GB/second bandwidth, and will operate within a 300 watt thermal design point (TDP) envelope. The Xeon Phi 5110P will offer additional performance at a lower power envelope.  Reaching 1.01 Teraflop/s and supporting 8GB of GDDR5 memory at 320GB/second bandwidth, it can achieve 225 watts TDP.  It is a passively cooled chip, delivering power efficiency that is ideal for dense computing environments.

With more than 75 percent of the November 2012 Top500 list choosing an Intel compute engine, the Xeon Phi coprocessor appeared in seven systems. These included the TACC Stampede effort with Dell, the Discover system at NASA Center for Climate Simulation and others.

The Intel Xeon Phi coprocessor 5110P is shipping immediately with general availability on January 28, 2013. The Xeon Phi 3100 product family will be available in the first half of 2013.

Vendor Support

SGI announced support of the Intel Xeon Phi coprocessor in fully managed and integrated solutions across its entire server product line. SGI has been an Intel Xeon Phi Validation Partner with early adoption of the Xeon Phi coprocessor on SGI server platforms at NASA Ames, COSMOS Consortium and The Genome Analysis Centre (TGAC), who are driving important advancements in scientific simulations and genomics. The Intel Xeon Phi is offered on the SGI UV 2000, SGI UV 20, Rackable and SGI ICE X platforms.

Professor Stephen Hawking said, "I am delighted that our new COSMOS supercomputer from SGI contains the latest many-core technology from Intel, the Xeon Phi coprocessors. With our powerful and flexible SGI UV2000 we can continue to focus on discovery, leading worldwide efforts to advance the understanding of our Universe."

Cirrascale announced the Cirrascale VB5400 blade server featuring support to handle up to eight of the newly released Intel Xeon Phi coprocessors. The company has developed a proprietary 80-lane PCIe switch, and integrates dual switches enabling it to support eight Intel Xeon Phi coprocessors in a single scalable system.

"As a valued partner, Cirrascale is once again bringing to market a solution that helps to further extends Intel's commitment for providing a revolutionary leap in discovery and innovation with the Intel Xeon Phi coprocessor," said Dr. Stephen Wheat, General Manager of High-Performance Computing at Intel. "The Cirrascale VB5400 with 8 Intel Xeon Ph™ coprocessors is an example of how this massively parallel processing solution can scale to enable High Performance Computing breakthroughs unlike anything previously seen."

Appro announced the availability and support of the Appro Xtreme Line of Supercomputers based on the Intel Xeon processor E5 family combined with the new Intel Xeon Phi coprocessors to address high performance, energy efficiency, programmability and manageability for scientific technical discoveries. Both Xtreme-X and Xtreme-Cool systems support two socket based on the Xeon processor E5 family combined with the new Xeon Phi Coprocessors. Appro was recently acquired by Cray.

“We are pleased to make available the next generation Appro Supercomputers based on Intel Xeon processors E5 family with a compatible option for Intel Xeon Phi coprocessors to be at the heart of our systems,” said Daniel Kim, CEO of Appro. “The combination of these processing technologies represents a leap forward in developing applications for parallel workloads while improving floating point performance, operational efficiency, and optimizations as new applications are introduced and updates to existing software packages are released.”

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.