PeakStream Ships Software Platform That Leverages New Multi-Core Processors

Company Provides Up to 10x Speed-Up for High Performance Computing Applications in Energy Exploration, Financial Services, Defense and Academic Research Markets: PeakStream, Inc., a leading software application platform provider for the high performance computing (HPC) market, today announced the immediate availability of its ground-breaking PeakStream Platform. The PeakStream software platform makes it possible to easily program new high performance processors such as multi-core CPUs, graphics processor units (GPUs) and Cell processors, converting them into radically powerful computing engines for exponentially increased application performance and decreased time-to-solution at a reduced cost. “The use of graphic processor units (GPUs) as application co-processor accelerators is very intriguing,” said Antony Jameson, Thomas V. Jones Professor of Engineering, Department of Aeronautics and Astronautics, at Stanford University. “Simplifying the programming environment for these devices will be critical for success. I'm looking forward to working with PeakStream in investigating applicability of GPU acceleration with my Computational Fluid Dynamics (CFD) codes.” The software platform, aimed at computationally-intense applications in markets such as energy exploration, financial services, defense signal processing and academic research, benefits any application requiring advanced computing horsepower, allowing engineers to leverage their existing skills and development tools to exploit the full potential of multi-core processors. “The potential to easily access the power of graphical processors is tantalizing for users of high performance computing in finance,” said John Liechty, Associate Professor of Marketing and Statistics at Penn State University. “Peakstream is clearly riding the wave of the future.” “The HPC industry’s response since our September launch has been overwhelming,” said Neil Knox, CEO and president of PeakStream. “Clearly, the marketplace feels there is a critical need to easily leverage the power of today’s industry-standard multi-core processors; and we are pleased that the PeakStream Platform is the first platform to offer a compelling solution to this challenge.” The PeakStream Platform features a stream programming model -- a data parallel approach to programming adapted for the distributed, hierarchical memories of the new generation of multi-core processor. PeakStream created this platform to enable the exponential performance gains and reduced development time prized by industries dependent on high performance computing. "Over the next five years, the HPC market will return to a more sustainable growth mode but will continue to see strong growth based on both the tendency of HPC applications to grow to fill the available computing power and the emergence of new applications in areas such as biosciences and national security,” said Earl Joseph, research vice president of high performance systems at IDC. “However, there is still the potential for ongoing double-digit growth in the smaller geographies. Vendors that are seeking strong growth rates are well advised to exploit these emerging markets. We believe that successful vendors will: Take advantage of technologies in innovative ways (e.g., by designing new types of systems around multi-core technologies).” Supported Platforms and Pricing The PeakStream Platform is offered in a Workstation Edition for single user application development and execution, and a Server Edition for cluster computing configurations. Pricing starts at $2,000 per node with volume discounts available for large clusters. Academic pricing is also available starting at $295. PeakStream also offers technical support, training and educational services. PeakStream recommends an AMD Opteron or Intel workstation or server equipped with a supported GPU. For a no charge evaluation of the PeakStream Platform, contact eval@peakstreaminc.com or call 866 4PK-STRM (866-475-7876). For more information on the PeakStream Platform, please visit: its Web site.