SCIENCE

SGI announced that the Institute of Fluid Science, Tohoku University in Sendai, Japan has begun operation of an SGI high performance computing (HPC) system. The HPC system consists of an SGI Altix UV 1000 operating at over 32 teraflops and delivering up to 4x peak performance over the previous system.

The new solution, added to other HPC systems already in place at the Institute, was designed and implemented by SGI Professional Services. It was configured with over 3550 cores of Intel Xeon processor X7560 series. It contains 10 terabytes (TB) of total memory with 4 TB of shared memory, and includes an integrated multi-screen Cave Automatic Virtual Environment (CAVE) 3-D visualization system. Applications such as Fluent, SGI OpenFOAM and in-house codes are utilized, enabling users to run AVS or Ensight to visualize the output data.

Institute of Fluid Science, Tohoku University is focused on high quality investigations on a wide range of studies. These include numerical simulation of sonic booms of supersonic aircraft, three-dimensional simulation of core-collapse supernovae, simulations and the design of intracranial stents for treating cerebral aneurysms and the associated blood flow analysis, meteorological forecast by downscaling simulations, measurement-integrated airport-area simulation for wake turbulence measurements, and numerical prediction of erosion of pipes in nuclear power plants due to liquid droplet impingement.

"We're pleased that Tohoku University's Institute of Fluid Science has been able to seamlessly integrate an SGI Altix UV 1000 into their existing infrastructure and bring it online at this time," said Nobuhiko Nakatsu, president of SGI Japan. "The results of their research will benefit communities around the world for years to come."

"Institute of Fluid Science, Tohoku University requires outstanding high performance computing technology to advance their research agenda," said Rajeeb Hazra, general manager of HPC at Intel. "By using the Intel Xeon processor X7560 series, the SGI Altix UV 1000 has the desired performance, memory capacity, and bandwidth to handle complex data intensive problems in fluid science. Intel is excited to help these scientists conduct collaborative research to move forward the state of the art."