Microsoft signals revolution in high performance computing

It’s not just that the company is, for the first time, offering software to run parallel, high-performance computing applications for complex engineering design and test, simulations and the like. It’s the fact that it all works seamlessly with the whole familiar Microsoft environment. Suddenly, the notion of a supercomputer on the desktop has become a reality. As Microsoft puts it: “Windows Compute Cluster Server 2003 accelerates customers’ time to insight by providing a reliable, HPC platform that is simple to deploy, operate, and integrate with existing infrastructure and tools.” General availability to customers is August, and evaluation versions are being provided to attendees of the Microsoft Tech•Ed 2006 conference, this week in Boston, USA. “High-performance computing technology holds great potential for expanding the opportunities within engineering, medical research, exploration and other critical human endeavors, but until now it has been too expensive and too difficult for many people to use,” says Bob Muglia, senior vice president of the Server and Tools Business at Microsoft. “We want HPC technology to become a pervasive resource — something that’s as easy to locate and use as printers are today.” And that’s what you’ll get. Windows Compute Cluster Server 2003 integrates out of the box with existing Windows infrastructures, and enables users to harness their existing development skills using Microsoft Visual Studio 2005. Beyond that, Microsoft has been working with the HPC community – particularly in the UK, but also around the world – and with strategic partners, to meet the requirements in terms of look and feel, interoperability, scaling and choice of compatible HPC applications. As part of that, Microsoft has spent millions of dollars in joint projects at academic institutions, including the University of Southampton, which has become one of 10 Microsoft HPC Institutes around the world, to help guide ongoing software research and product innovation. The system has already proven itself among early-adopter users for vehicle design and safety improvements in the automotive sector, for oil and gas reservoir simulation and seismic processing, and for simulations of enzyme catalysis and protein folding with life sciences users. Such users include BAE Systems, Queen’s University Belfast, Cornell University’s Computational Biology Service Unit, the National Center for Atmospheric Research, Northrop Grumman, Petrobras (Brazil), Tokyo Institute of Technology’s Global Scientific Information and Computing Center, the University of Cincinnati’s Genome Research Institute and Virginia Tech’s Computational Bioinformatics and Bioimaging Laboratory. Defence and aerospace firm BAE Systems, for example, says it migrated key engineering platforms to the new Microsoft platform, along with Windows Workflow Foundation and SQL Server, and is now speeding up the design process and improving product performance, while also significantly cutting the cost and investment required to run its high performance computers. “Simplifying our fluid dynamics engineering platform will increase our ability to bring solutions to market and reduce risk and cost to both BAE Systems and its customers,” says Jamil Appa, BAE Systems group leader for Technology and Engineering Services. The big question: how long will it take the computing, engineering, technology and business hardware, software and services community to catch up? Not long. Microsoft is already working with a long list of partners to bring the HPC transformation about. By the end of 2006, says the firm, we can expect release solutions from IT firms including: Abaqus, Absoft, AMD, ANSYS, BioTeam, Broadcom, Cisco Systems, Dell, ESI Group, Fluent, Fujitsu, Hitachi, HP, IBM, Intel, the MathWorks, Mellanox, MSC Software, Myricom, NEC, Parallel Geoscience, Platform Computing, the Portland Group, Schlumberger, SilverStorm Tyan Computer, Verari Systems, Voltaire and Wolfram Research. And on the hardware side, the time is already right, with multi-core processors, standards-based, high-speed interconnects and x64 (64-bit x86 architecture) devices also mainstream and at low cost per node. Evaluation versions are available now from http://www.microsoft.com/hpc, and Microsoft says the price will be around $469 per node.