Fornax supercomputer for WA university

A $4 million terascale research supercomputer, funded under the Australian Government’s Super Science Initiative, will help researchers at the University of Western Australia to carry out leading-edge computational research. Named Fornax, Latin for ‘furnace’, it is the second of three supercomputers commissioned as part of the $80 million Pawsey Centre project.

“As part of Pawsey Centre, Fornax is at the heart of Australia and New Zealand’s successful bid to co-host the Square Kilometre Array radio telescope,” said Senator Chris Evans, Minister for Science and Research.

“Operated by expert researchers and technical staff, Fornax offers a supercomputing resource which is designed to take in masses of seismic and astronomical data, and forge complex computations out of them.

“It gives Australian researchers access to the kind of computing power that is critical to astronomy signal processing needed for the international SKA project.

“This is the second of two forerunners to the more powerful petascale Pawsey Centre supercomputer system being installed in 2013. Fornax and its partner Epic are helping researchers to develop the expertise needed to get the best out of the supercomputer when it comes online.

“With these systems, we’re building world-class supercomputing facilities, and expertise to match, that will put Australia at the forefront of research in radioastronomy, the geosciences and other high-end computational research.”

Senator Evans said the Pawsey Centre was a pivotal component of the government’s investment of almost $9 billion in science, research and innovation since 2007.

“We’re committed to ensuring Australia plays a lead role in the international science community, including in radioastronomy research, through investments like the Pawsey Centre,” Senator Evans said.

Fornax is operated by iVEC, a collaboration between CSIRO and University of Western Australia, Murdoch University, Curtin University and Edith Cowan University.

The system contains 6.9 terabytes of RAM, 672 terabytes of storage and 1152 CPUs, operating across 96 nodes. At peak performance, the system is capable of performing 62 teraflops, or 62 trillion (62,000,000,000,000) operations per second.