SGI has announced the availability of NVIDIA Tesla 20-series GPUs, based on the latest generation Fermi architecture, for SGI's GPU compute solutions. Delivering application-specific acceleration, SGI's GPU computing portfolio ranges from the deskside to the supercomputer with products optimized to address the most computationally-intensive environments in the world. As a leader in acceleration and design-to-order solutions, SGI is uniquely positioned to provide NVIDIA Fermi integration, bringing customers hybrid solutions of traditional microprocessors and accelerators that solve critical technical computing problems.

SGI will showcase one of its NVIDIA Tesla GPU-enabled platforms, Altix UV, at the GPU Technology Conference, September 20 -- 23, 2010, in San Jose, California, in booth number 84.

"With NVIDIA Tesla GPUs integrated into the SGI product portfolio, customers can now deploy solutions that are optimized to address their most challenging technical computing problems," said Andy Keane, general manager, Tesla business at NVIDIA.

SGI's unique GPU compute solutions immediately available with NVIDIA Fermi include:

• Deskside High-Performance Computing with Octane III -- Offers the capabilities of a high-performance cluster with the portability and usability of a workstation. It is available as a single- or dual-node, dual-socket graphics workstation and an integrated cluster with support for the fastest NVIDIA graphics and GPU compute cards

• Workgroup to Enterprise with SGI Altix XE and Rackable Servers -- Leveraging the winning combination of the latest Intel Xeon processor architecture, NVIDIA Quadro professional graphics and Tesla 10 and 20-series solutions, these servers deliver top value and performance.

• Supercomputing with Altix UV -- With up to eight NVIDIA Quadro and Tesla 20-series GPUs combined in a single-system, the Altix UV platform brings GPU acceleration to a new class of performance solutions in chemistry, homeland defense, fluid dynamics and biosciences.

• Supercomputing with Altix ICE -- For customers who want to manage large scale-out HPC environments that include GPUs, the SGI Altix ICE 8400 platform offers the ability to integrate service nodes containing GPUs into a dual-plane, high-bandwidth, low-latency InfiniBand network.

"SGI leads the industry in delivering application-specific acceleration with partners such as NVIDIA," said Bill Mannel, vice president of product marketing at SGI. "SGI is excited to continue our relationship with NVIDIA, integrating Fermi into our GPU portfolio for best-in-class graphics and GPU options. Additionally, we look forward to integrating NVIDIA Tesla GPUs into future offerings, including 'Project Mojo,' our hybrid computing platform."

Mr. Mannel will be presenting "What If You Had a Petabyte of Memory and/or a Petaflop of Compute?" at the GPU Technology Conference on Wednesday, September 22, at 10:00 a.m. PT. This session will explore application spaces where GPU compute coupled with very large shared memory architectures and/or petaflops of compute are allowing new science and business questions to be addressed.

In addition to support for NVIDIA Tesla GPUs, SGI recently coupled its hardware solutions with a team of application engineering and professional services experts able to tune customer applications for GPU computing. Because of this expertise, the Irish Centre for High-End Computing (ICHEC) chose to work exclusively with SGI to enable institutions and businesses to accelerate the development of applications that utilize GPU compute systems.