Supercomputer maker Cray Inc. today launched its next generation high-end supercomputing systems -- the Cray XC30 supercomputer. The system takes advantage of Intel's Aries interconnect technology to offers greatly improved internal bandwidth.
Previously code-named "Cascade," the Cray XC30 supercomputer is the company's most-advanced high performance computing system ever built. The Cray XC30 combines the new Aries interconnect, Intel Xeon processors, Cray's software environment, and new power and cooling technologies to create a production supercomputer that is designed to scale high performance computing (HPC) workloads of more than 100 petaflops.
The Cray XC30 supercomputer has been engineered to meet the real-world performance challenges of HPC users. Cray's new high-end system features the new HPC-optimized Aries system interconnect; a new Dragonfly topology that frees applications from locality constraints; an innovative cooling system that utilizes a transverse airflow; the next-generation of the scalable Cray Linux Environment that also supports a wide range of ISV applications; Cray's HPC optimized programming environment; and the ability to handle a wide variety of processor types including the Intel Xeon processors.
The Cray XC30 will utilize the Intel Xeon processors E5-2600 product family and with these Intel processors, Cray XC30 systems can scale in excess of one million cores. Additionally, future versions of the Cray XC family of supercomputers will be available with the new Intel Xeon Phi coprocessors and NVIDIA Tesla GPUs based on the next-generation NVIDIA Kepler GPU computing architecture.
Each node in the new systems gets its own 48-port Aries interconnect. Aries performs the duties of both a network card and a router, providing each node direct access to every other node in the system.
Thanks to the new Aries interconnect technology, each node is capable of 120 million gets/puts per second, which compares to 30 to 40 million gets/puts per second for the previous generation of Cray systems.
Cray sold its interconnect hardware development program, including much of the intellectual property behind Aries, to Intel in May for US$140 million.
The XC architecture also has a new network topology called Dragonfly. Dragonfly is considered a direct network topology, one that requires fewer optical links and no external top switches.
Each cabinet will consist of three chassis, with each chassis having 16 computer blades, and each blade comprising four nodes. The system can be built with as many as 92,544 nodes, each with up to 128GB of DDR3 memory.
Early shipments of the Cray XC30 are starting now, and systems are expected to be widely available in first quarter of 2013.