AMD awarded extreme-scale computing grant to solve problems of climate science

The US Department of Energy (DOE) this week awarded Advanced Micro Devices (AMD) a $3.1 million (£1.9 million) grant to research computer interconnect technology as part of a DOE program to develop extreme-scale computing called DesignForward, the chip maker said.

AMD also received a monetary award in 2012 as part of an earlier iteration of the DOE's Extreme-Scale Computing Research and Development Program called FastForward. This week's grant is an expansion of the two-year research grant the company got for the FastForward program.

"This award will enable AMD to continue important research to help realise the enormous potential of extreme-scale computing. We're excited that our expertise in processors, memory, and interconnect technology will be leveraged for this groundbreaking research in large-scale systems," Alan Lee, AMD corporate vice president of Research and Advanced Development, said in a statement.

While the FastForward program concerns processor and memory technologies for extreme-scale computing, DesignForward participants are working on the development of interconnect architectures and making data transfer capabilities within and between systems more efficient.

The DOE's extreme-scale computing efforts aim to produce "exascale" supercomputers "capable of performing computation 30 times faster than today's fastest computers, with only slightly higher power utilisation," AMD noted. The DesignForward program is being funded by the DOE's Office of Science and the US National Nuclear Security Administration (NNSA).

To build a true exascale supercomputer, "leaps in research and development need to be advanced in all aspects of computing," the company said.

Lee said AMD's expertise in developing heterogeneous computing platforms that leverage multiple processor and chipset architectures is a good fit for the DesignForward effort. AMD, along with ARM, TI, and other semiconductor firms formed the Heterogeneous System Architecture (HSA) Foundation in June 2012 to promote the "industry-wide advancement of standard, open systems," he said.

"The DOE's strategic plan seeks to address the nation's most pressing scientific challenges by advancing simulation-based scientific discovery made possible by the world's highest performing exascale computers. Among its priorities are: advancing medical research, alternative energy and energy storage materials, industrial design and more precise models for climate and weather change," Lee said.

Earlier this year, Horst Simon, the Deputy Director at the Lawrence Berkeley National Laboratory's National Energy Research Scientific Computing Centre (NERSC) bet $2,000 (£1,240) that the world wouldn't reach exascale levels of computing before 2020.

Will AMD put him out of pocket?

Image: (flatworldsedge; Berkeley Lab)