AMD Receives Funding For Exascale Computing R&D

The funding will consist of $9.6 million for processor-related research and up to $3 million for memory-related research, AMD said. The DOE is offering the funding as part of a larger extreme-scale computing R&D initiative called FastForward, which it launched alongside the National Nuclear Security Administration (NNSA) in April.

The DOE will grant approximately $60 million in total funding over the next two years to manufacturers in the processor, memory and storage markets.

[Related: The 10 Coolest CPUs, GPUs Of 2012 (So Far) ]

"To prepare for the next phase of extreme-scale computing, NNSA and DOE Office of Science are taking a proactive step in jointly making strategic investments in key areas such as processor, file storage and memory technologies with AMD and others," said Thuc Hoang of the DOE’s NNSA, in a statement. "A key to successfully developing next-generation HPC capabilities is bringing together the know-how and best minds of industry leaders and national labs to work on this grand challenge."

Cloud storage vendor Whamcloud also joined the FastForward program this week, with a specific focus on storage and I/O research.

Chip makers and government agencies have been focusing on exascale computing, as it is predicted to yield significantly faster processing times for data-intensive applications, such as those used in health care and defense. When the exascale achievement is reached, supercomputers will run roughly one thousand times faster than the highest-end systems used today.

Joining the FastForward program is AMD's latest move to accelerate the development of new technologies specific to exascale computing. In November, the Sunnyvale, Calif.-based chip maker hosted a roundtable with a number of executives in the high-performance computing space, including Margaret Williams, senior vice president of HPC Systems at Cray, and Charles Wuischpard, president and CEO of Penguin Computing, to explore ways in which the cloud and heterogeneous computing could be used to help reach the exascale milestone.

Fellow chip maker Intel also has stressed the importance of HPC, unveiling in June its first-generation Xeon Phi family of co-processors, code-named Knights Corners, that it said provides the foundation for enabling exascale computing by 2018.

PUBLISHED JULY 11, 2012