Case Study Reveals Cray CX1 System Increases Delivery of Complex Analyses at National Supercomputing Center for Energy and the Environment

SEATTLE, WA -- (Marketwire) -- 06/24/09 -- Global supercomputer leader Cray Inc. (NASDAQ: CRAY) today announced that the National Supercomputing Center for Energy and the Environment (NSCEE) tested the Cray CX1(TM) deskside supercomputer running Microsoft Windows HPC Server 2008 and found the system is 46 percent faster in the delivery of complex analyses, resulting in greater researcher productivity. The results were reported in a newly released case study published by Cray and Microsoft's high performance computing (HPC) groups.

A full-service supercomputing facility located at the University of Nevada in Las Vegas, the NSCEE maintains the most advanced HPC systems in Nevada. Although the NSCEE maintains a range of computing resources, the center sought a flexible, affordable option that researchers could easily use without needing to access one of the center's large supercomputers.

The NSCEE was introduced to a new computer, the Cray CX1 deskside supercomputer with Windows HPC Server 2008 and decided to conduct a series of benchmarking sessions to determine how the performance of the Cray CX1 system with 32 cores compared with traditional 32-core HPC solutions.

In October 2008, the NSCEE deployed the Cray CX1 system with Windows HPC Server 2008 and soon began running tests. The resulting case study with NSCEE highlighted the following results:

--  Faster performance. The NSCEE discovered that benchmarking results
    indicated the Cray CX1 system with Windows HPC Server 2008 ran 46 percent
    faster than a 32-core commercial, off-the-shelf cluster.
--  Greater productivity. NSCEE researchers could increase productivity
    and quickly solve problems by getting results in days, rather than weeks.
--  Ease of use.  With a familiar Windows user interface paired with a
    supercomputing system that is small, quiet and purpose-built for the office
    environment, NSCEE scientists didn't have to spend valuable research time
    learning a new HPC system.
--  Cost savings.