Supercomputer Accelerates Pace of Health Research

The University of Manchester (United Kingdom) is financing a new supercomputer system that can analyze bigger and more complex sets of information.

The Health eResearch Center (HeRC; Manchester, United Kingdom) will allow researchers from eight leading research universities access to the new Silicon Graphics International (SGI; Milpitas, CA, USA) UV2 supercomputer. With an objective of using health data to improve understanding of diseases and develop better treatments and care, HeRC will merge health research and clinical practice with academic disciplines of computer science and statistics. HeRC, based in northern England, is one of four centers across the UK with partners in Scotland, Wales, and London, collectively known as The Farr Institute.

The eight universities (Durham, Lancaster, Leeds, Liverpool, Manchester, Newcastle, Sheffield, and York) collaborating together as the N8 Research Partnership already have access to a High Performance Computer (HPC) based at the University of Leeds that is capable of a peak performance of 110 trillion operations per second. Housing the new HeRC supercomputer alongside the existing facility will allow the research partnership and industry partners to speed up even further the pace of health research by allowing scientists to process data in quicker and more efficient ways.

“We are developing new ways to analyze public health information to improve our understanding of common diseases,” said John Ainsworth, MD, deputy director of HeRC. “This advanced computer system will lead to much more powerful research being conducted and enhance the speed at which researchers can produce results.”

“This is an exciting development for N8 HPC,” added Professor Chris Taylor, N8 HPC co-director. “It allows us to support a nationally important initiative in e-Health, providing a cost-effective solution by building on the existing N8 HPC infrastructure. It also extends the capability of N8 HPC for all users—a win-win.”

The SGI UV2, which runs on Intel Xeon E5 processors, doubles the number of cores (up to 4,096 cores) and quadruples the amount of coherent main memory (up to 64 terabytes) from the previous UV generation, scaling up to eight petabytes of shared memory at a peak I/O rate of four terabytes per second (14 PB/hour). For example, it could ingest the entire contents of the US Library of Congress print collection in less than three seconds.

Related Links:

Health eResearch Center
Silicon Graphics International
N8 Research Partnership