WALTHAM, Mass., Nov. 18, 2008 – Interactive Supercomputing Inc. (ISC) launched a new version of its Star-P® software that features a wide array of new or improved function libraries to make High Performance Computing (HPC) applications run faster and users more productive.
The new version will also expand its platform support to include Microsoft Windows® HPC Server 2008 and Intel® Cluster-Ready systems, and incorporates more than 250 software improvements based on customer recommendations – including enhanced parallel file I/O and capability to support the security needs of our enterprise customers.
Star-P empowers scientists, engineers and analysts to create algorithms and models on their desktops using familiar mathematical tools - such as MATLAB® and Python - and then run them instantly and interactively on parallel computers such as servers, clusters, nodes on a grid or On-Demand resources. It eliminates the need to re-program applications to run on parallel systems, making problem-solving faster and users more productive.
Star-P 2.7 includes a clean sheet reimplementation of function libraries which concentrated on accuracy, performance and ease of use. The libraries include hundreds of individual computing functions in key computational domains such as statistics, optimization, image and signal processing. The new libraries have numerous improvements in precision, and speed when performing compute-intensive operations performed on large, distributed data sets.
“While Star-P is well known for ease of use and exceptional parallel performance, our customers also expect the best in numerical algorithms. And as a numerical analyst, issues of accuracy are of the utmost importance to me and our customers,” said Alan Edelman, ISC’s chief science officer. “The ground-up reimplementation of functions, which was a major investment, allows us to provide exceptional performance and functionality in Star-P 2.7 and future releases.”
Adding Windows HPC Server 2008and Intel Cluster Ready platform support in Star-P 2.7 makes it easier for customers to buy and deploy HPCs without the prohibitive cost and programming complexities associated with traditional cluster computing. Windows HPC Server 2008 combines the power of the Windows Server platform with out-of-the-box functionality to improve the productivity and reduce the complexity of HPC environments. Intel Cluster Ready support means that Star-P 2.7 will be compatible with a wide range of Intel-based multi-core servers and clusters that are Intel Cluster Ready certified.
The new version is being demonstrated today on a Cray CX-1 supercomputer running Windows HPC Server, an Intel-based Dell cluster linked with Mellanox’ Infiniband switch architecture, and an HP BladeSystem C3000. The demos can be seen at ISC booth #641 at the SC ’08 conference in Austin, Texas.
“Our experience has shown that the road to easy parallel performance is paved with both sophisticated architecture and countless low-level features which both protect the programmer from errors and automate complicated data distribution functions,” said David Rich, ISC vice president of marketing. “We’re very happy with the large number of improvements we’ve made and believe our customer’s will see benefit from increased productivity and performance.”
Star-P 2.7 is in beta test now and will be generally available in December 2008.
About Interactive Supercomputing
Interactive Supercomputing (ISC) launched in 2004 to commercialize Star-P, an interactive parallel computing platform. With automatic parallelization and interactive execution of existing desktop simulation applications, Star-P merges two previously distinct environments - desktop computers and high performance servers - into one. Based in Waltham, Mass., the privately held company markets Star-P for a range of biomedical, financial, and government applications. For additional information, visit www.interactivesupercomputing.com.
# # #
Star-P is a registered trademark of Interactive Supercomputing Inc. All other trademarks mentioned herein are the property of their respective owners.