GOVERNMENT

Global shared-memory architecture of the SGI Altix 4700 fuels climate research: March 1, 2007, to March 1, 2008, marks the fourth International Polar Year (IPY) -- a large scientific program focused on research in both polar caps, the Arctic and the Antarctic, to recognize the strong links these regions have with the global climate changes. This year's activities mark China's first IPY efforts, and the Polar Research Institute of China (PRIC) selected compute technology from SGI to enable this important scientific research mission. "Our research is focused on how the Earth's Polar Regions affect global climate systems," said Dr. Sun, head of the Oceanographic Polar Region Research Lab. "With the study of the ice-sea model, we will predict the global climate changes as well as low and medium latitude climate trends. We are not computer science experts, but we found SGI's supercomputer very easy to deploy to serve our research goal. SGI's engineers are very professional to set up not only the hardware, but also the software environment and application code. SGI's turnkey solution is expected to accelerate our research and time to insight significantly." PRIC researchers are running a host of climate models which benefit significantly from the 64GB of shared-memory of their new SGI Altix 4700 server. Built using SGI's powerful scalable shared-memory architecture and featuring a high-performance, industry-standard 64-bit Linux environment, the SGI Altix 4700 enables all processors direct access to the global shared- memory for optimal performance on big data problems and ease of programming. Among the climate models being used by PRIC are:

• GFDL Modular Ocean Model (MOM4), a 3-dimensional, z-coordinate, primitive equation ocean circulation model
• Princeton Ocean Model (POM), a sigma coordinate, free surface, ocean model, which includes a turbulence sub-model
• NCAR's Spectral Transform SWM (Shallow Water Model), a parallel algorithm that solves the nonlinear shallow water equations on a rotating sphere using the spectral transform method
• Atmosphere-Ocean Model (AOM), designed at GISS for climate predictions at decade to century time scales

"SGI applauds the remarkable work of these dedicated researchers in unlocking the secrets to global climate change, said Alex Lee, country manager of SGI Greater China Region. "As more and more data are being generated from multi-disciplined research about the Polar regions, the integrated high performance and scalability of SGI Altix systems continue to help scientists and engineers achieve what just a short time ago was considered impossible. The achievements of these scientists illustrate that SGI truly delivers Innovation for Results." "This system serves almost all the disciplines of our research area," said Mr. Zhu Jiangang, Director of Polar Information Center, who serves different labs within the institute added, "Besides the oceanographic research labs, some other labs, like upper atmosphere physics lab also migrated their own developed code from PC clusters to this platform with very positive results. They were able to significantly improve their productivity." PRIC's new SGI Altix 4700 system powered by 32 Intel Intanium 2 processors with 64GB globally shared memory, and running Novell's SUSE Linux Enterprise Server 10, was installed in May.