Cray Adds Intel Xeon Phi Processor to Flagship Line of Supercomputers

Print Friendly, PDF & Email

crayToday Cray introduced new performance breakthroughs that will provide customers with the fastest Cray XC supercomputers and Cray Sonexion storage systems to date.

The new Intel Xeon Phi processor family, formerly code named “Knights Landing”, is now available in the Cray XC series of supercomputers, offering peak performance of over half-a-petaflop per cabinet – a 2X performance boost over previous generations. Cray also unveiled the new Cray Sonexion 3000 scale-out Lustre storage system, which features capacity- and performance-optimized configurations. In a performance-optimized configuration, the Cray Sonexion system delivers nearly 100 gigabytes-per-second of performance in a single rack.

“Our customers are taking on increasingly complex computational problems that are expanding the boundaries of supercomputing and storage performance capabilities,” said Ryan Waite, Cray’s senior vice president of products. “We partner closely with our customers to understand their unique requirements and deliver new systems that deliver peak performance. For many of our customers, Intel Xeon Phi processors and Lustre parallel file systems are critical components of their supercomputing infrastructure. Our close collaboration with Intel helps to ensure our Intel Xeon Phi processor-based solutions scale to the most demanding performance requirements and our close partnership with Seagate helps scale Lustre to new levels of performance and stability.”

The Cray XC supercomputer is an adaptive supercomputing system that supports different processing and storage technologies in the same architecture. It also features a fully-integrated software ecosystem that leverages Cray’s vast expertise in many-core and multi-core computing environments. With the new Intel Xeon Phi processor, Cray XC customers can take advantage of the Cray Programming Environment to optimize applications for best performance and memory usage. The software stack in the Cray XC system also includes enhancements to help customers optimize code and best utilize the new high bandwidth memory integrated in the Intel Xeon Phi processor.

Cray is a leader in deploying supercomputers with the new Intel Xeon Phi processor, and several top supercomputing centers have already signed large contracts for Cray XC supercomputers with Intel’s new processor. These early systems include the “Cori” system at the U.S. Department of Energy’s (DOE) National Energy Research Scientific Computing Center (NERSC); the “Trinity” system at the National Nuclear Security Administration; and the “Theta” system at the DOE’s Argonne Leadership Computing Facility at Argonne National Laboratory.

As a foundational element of the Intel® Scalable System Framework, the Intel Xeon Phi processor is optimized for scalability, compute parallelism, and memory bandwidth,” said Charles Wuischpard, vice president Data Center Group, general manager of the High Performance Computing Platform Group, Intel. “Cray XC systems feature highly innovative supercomputing technologies, and we are excited that our collaboration with Cray continues to enable researchers and scientists to achieve breakthrough innovations and discoveries.”

Designed to match increases in compute performance, the Cray Sonexion 3000 system builds on Cray’s expertise in performance-engineered, parallel storage systems for supercomputing and big data analytics. Compared to conventional Lustre solutions, the Cray Sonexion 3000 system is designed to scale more efficiently, deliver up to 38 percent more real-world throughput per rack unit, and reduce total-cost-of-ownership by up to 25 percent. The new storage system is pre-integrated and fully tested, and Cray offers a single point of support for all hardware and software.

Many organizations now need deep expertise to develop scalable, parallel applications that can extract the full value of high performance, scale-out storage and compute resources,” said Nik Rouda, Senior Analyst, Enterprise Strategy Group “Cray is closing the gap, bringing supercomputing capabilities to big data and analytics in the enterprise.”

Sign up for our insideHPC Newsletter