CHIP DESIGNER Nvidia has announced that the Barcelona Supercomputing Center (BSC) is developing a hybrid supercomputer using its CPUs and GPUs.
The system will employ Nividia's Tegra ARM CPUs combined with its Cuda GPUs. The BSC's aim is to build a system, in the not too distant future, that is between two and five times better in terms of energy efficiency than today's most efficient systems. Nvidia is showing off the design at this week's SC11 conference in Seattle.
"In most current systems, CPUs alone consume the lion's share of the energy, often 40 percent or more," said Alex Ramirez, leader of the Mont-Blanc Project. "By comparison, the Mont-Blanc architecture will rely on energy-efficient compute accelerators and ARM processors used in embedded and mobile devices to achieve a four- to 10-times increase in energy-efficiency by 2014."
BSC said its ultimate goal is deliver exascale-level performance but using between 15 and 30 times less power than current supercomputer architectures. The project will explore next generation high performance computing architectures and develop applications for the technologies.
Nvidia also announced plans to develop a hardware and software development kit (SDK) to support demand for similar ARM initiatives. The kit will apparently feature a quad-core Nvidia Tegra 3 processor with a discrete GPU and will be available in the first half of next year. µ
Sign up for INQbot – a weekly roundup of the best from the INQ