New life-sciences supercomputer called Koronis expected at University within the next month

The $3.7-million grant, awarded this summer, allowed the University to buy the supercomputer with 1,152 core processors and 3.1 terabytes of memory.

by Frank

A grant awarded to the University of Minnesota this summer will result in a new computing system for the University’s Supercomputing Institute for Advanced Computational Research.

The new supercomputer called Koronis is expected to arrive on campus in the next month and will allow University researchers to handle large amounts of data in life-sciences research.

More than 30 researchers from the University collectively applied for the National Institutes of Health grant; the researchers represented various areas of study including chemistry, computer science, biomedicine, chemical engineering, medicinal chemistry and psychology.

The $3.7 million grant made the purchase of Koronis possible.

Koronis, built by supercomputer manufacturer SGI, is composed of 1,152 core processors, and 3.1 terabytes of memory — 3,100 gigabytes.

A typical laptop at Best Buy will have a dual core processor and roughly two gigabytes of memory, said H. Birali Runesha, director of Scientific Computing and Applications at the University.

The type of memory and the size of the disk space is what sets Koronis apart from other supercomputers, said Bill Hellriegel, assistant to the director of the University’s Supercomputing Institute.

Koronis uses a shared memory system, as opposed to a cluster memory system, which allows thousands of the computer’s processors to access the same memory.

Koronis will have three-quarters of a petabyte of raw disk memory — roughly 780,000 gigabytes of disk space and almost 4,000 times the average of a typical computer.

Other supercomputers would not be able to handle the enormous disk space requirements needed to run certain programs, Runesha said.

“We expect the system to benefit a lot of people and facilitate their research,” Hellriegel said.

“Historically, science was observation, then it moved into theoretical, and now you’ve got the computational aspect,” Brian Ropers-Huilman, director of Systems Administration and Technical Operations, said. “Having the unique ability for our researchers to handle these massive data sets is a huge advantage [for the University].”