Swinburne University of Technology's new $3 million GPU Supercomputer for Theoretical Astrophysics, 'gSTAR', has been officially launched.

 

gSTAR represents the next generation of supercomputing, and is one of only six such machines in Australia and among the top 200 in the world.

 

Astronomers have employed the same graphics processing units (GPUs) used in personal computers and game consoles to build the 120-teraflop supercomputer capable of modelling entire universes.

 

Director of the Swinburne Centre for Astrophysics and Supercomputing, Professor Warrick Couch, said gSTAR is set to accelerate the rate of science and knowledge discovery at Swinburne and is at least 10 times faster than its predecessor.

 

"This means that the three years of data processing that culminated in early 2011 in the discovery of a diamond planet in the Milky Way galaxy could now potentially be done in one week," Professor Couch said.

 

"Traditional supercomputers rely on central processing units (CPUs) to process information, whereas gSTAR uses GPUs that can perform numerous high-speed tasks simultaneously," he said.

 

The system is supported by a quad data rate (QDR) InfiniBand network that transfers data to GPUs and large memory nodes, each with 512 gigabytes of memory, at 1000 times the speed of a wireless internet connection.

 

Professor Couch said gSTAR is poised to receive a deluge of data from an emerging generation of telescopes equipped to generate a million billion bytes of information in a single night.

 

"GPUs will make a major contribution to processing data from new optical and radio telescopes, such as SkyMapper in NSW and the Australian Square Kilometre Array Pathfinder in Western Australia," Professor Couch said.

 

The massive jump in capability has involved the investment of more than $1.9 million by Swinburne and $1.04 million from the Federal Government's Education Investment Fund.

 

Silicon Graphics International (SGI) has supplied the hardware for gSTAR. SGI Australia and New Zealand General Manager, Nick Gorga said Swinburne's scientific discovery profile has been raised as a result of hosting gSTAR.

 

 

"This additional capacity gives Swinburne a distinct technological advantage and will rank this system amongst the highest within the south Asia region," Mr Gorga said.

 

gSTAR uses 636 Intel computer processing units (CPUs) and 121 NVIDIA graphics processing units (GPUs).  It is contained in 11 cabinets on the fifth floor of the George Swinburne building at Swinburne's Hawthorn Campus, and will meet the supercomputing needs across a number of faculties.