Obama Looks to Accelerate HPC Innovation

By Jeffrey Burt  |  Posted 2015-07-31 Print this article Print
data center

The president's executive order creates a federal body to coordinate government efforts to push development of the first exascale computer.

The Obama Administration wants to use the resources of the federal government to accelerate innovation in the country's high-performance computing market and to drive the development of exascale compute systems in the United States.

President Obama this week signed an executive order establishing the National Strategic Computing Initiative (NSCI) to coordinate federal government efforts with those from public research institutions and the private sector to create a comprehensive, long-term strategy for the development of high-performance computing (HPC) systems.

In his executive order, Obama noted that the United States has for six decades driven research and development around new computing systems that offer increasingly faster application performance. The president's order also noted that the innovation on and deployment of these machines has contributed greatly to government, industry and academia.

That momentum needs to continue and can be helped by taking a "whole of government" approach, he wrote.

"Maximizing the benefits of HPC in the coming decades will require an effective national response to increasing demands for computing power, emerging technological challenges and opportunities, and growing economic dependency on and competition with other nations," the president wrote. "This national response will require a cohesive, strategic effort within the Federal Government and a close collaboration between the public and private sectors."

Government agencies have been pushing the industry toward exascale computing. For example, the Department of Energy (DoE) for several years has been giving money to tech companies to build faster supercomputers for its various laboratories. Most recently, the DoE in April announced it was awarding $200 million to Intel and Cray to build an 180-petaflop supercomputer for its Argonne National Laboratory in Illinois to be delivered in 2018.The system is part of the DoE's Collaboration of Oak Ridge, Argonne and Lawrence Livermore (CORAL) initiative, a $525 million program launched in November 2014 to build supercomputers for each of the national labs that are five to seven times faster than current systems and will help the United States in the push toward exascale computing.

The Obama Administration is taking a long-term view of the effort. Among its five objects, the the executive order not only talks about accelerating the development and deployment of an exascale system that can deliver 100 times the performance of current 10-petaflop systems across a range of applications, but also—over the next 15 years—creating the technology needed for future HPC systems after current semiconductor technologies hit their limits and Moore's Law is exhausted.

The other objects include the continued merging between technologies used for modeling and simulation and those used for data analytics, employing a holistic approach to address a broad range of technological factors—from networking and workflow to downward scaling, foundational algorithms and software, accessibility and workforce development—and establishing an ongoing public-private collaboration to ensure that R&D advances are shared between the government, academia and private sectors.

The United States remains the dominant HPC player in the world. The country housed the most systems on the latest Top500 list of the world's fastest supercomputers, released in July, with 233, and five of those were in the top 10. However, the number is declining: in November 2013, the United States had 265 systems on the Top500.

At the same time, China's Tianhe-2 supercomputer once again was at the top the list of the world's fastest systems, a position it now has held for more than two years since overtaking the U.S.-based Titan supercomputer housed at the Oak Ridge National Lab and based on Cray's XK7 systems that are powered by Opteron processors from Advanced Micro Devices and Nvidia Tesla K20x GPU accelerators.

In a post on the Nvidia blog, Ian Buck, vice president of the company's Accelerated Computing business unit, wrote that to get to exascale computing, the systems will need to rely on a range of modern technologies, including GPU accelerators, for greater performance and energy efficiency.

"The payoff from machines efficient enough to reach exaflop speeds could be enormous," Buck wrote. "An exaflop computer would have the potential to provide unprecedented insight into the workings of the human brain. Or lead to breakthroughs in personalized medicine. … Building ever faster machines will require rethinking the way these machines are built."


Submit a Comment

Loading Comments...
Manage your Newsletters: Login   Register My Newsletters

Rocket Fuel