Science Committee Discusses Exascale Computing Research and Development

A May 22 hearing of the House Science, Space and Technology’s Subcommittee on Energy demonstrated bi-partisan enthusiasm for advanced computing technology. Members were interested in learning about developments and challenges faced by researchers as they relate to exascale computing. They acknowledged that high performance computing has played a role in many scientific disciplines and that it has revolutionized how many economic sectors operate.

During the hearing Representative Randy Hultgren (R-IL) exemplified that support for computing research as he described his discussion draft of The American High-End Computing Leadership Act. The bill would modify the Department of Energy (DOE) High-End Computing Revitalization Act of 2004 to authorize increased funding for the DOE Office of Science to develop exoscale computing systems at national laboratories. The bill establishes national laboratory – industry partnerships that would allow industry, institutions of higher education, and Federal agencies to work jointly on the development of exascale computer systems. Furthermore, the bill calls for the DOE to submit an integrated strategy and program management plan to Congress which would outline the execution of the Exascale Computing Program. That plan would address the prioritization of programs during times of tight fiscal climates.

“Exascale computing represents an exciting new world of science for our nation. The applications for the next generation of super computers are vast; and by solving the challenges in basic science to develop these systems, we will ensure our nation’s long-term innovation and competitiveness,” said Hultgren.

While highlighting the importance of high performance computing, Subcommittee Chairwoman Cynthia Lummis (R-WY) noted that “these advancements ultimately trace back to Federal investments in basic research that provided the foundation for most of today’s computing technologies. From the first megaflop supercomputers of the 1960s, Federal investments have led the push across each landmark thousand-fold speed barrier—to gigaflops, teraflops, and petaflops.” She went on to describe the role of the Advanced Scientific Computing Research program at DOE as a part of DOE’s overall support for facilities including the National Energy Research Scientific Computing Center.

Building on the positive impact of computing centers, Lummis emphasized that “it is important we take time to plan and budget thoroughly to ensure a balanced approach that ensures broad buy-in from the scientific computing community. The Federal government has limited resources and taxpayer funding must be spent on the most impactful projects. We need to ensure DOE efforts to develop an exascale system can be undertaken in concert with other foundational advanced scientific computing activities.”

Ranking Member Eric Swalwell (D-CA) was eager to learn more about DOE computing partnerships with industry and national laboratories. He called attention to the vast examples of industrial and academic research that has benefited from advanced computing including energy research, pharmaceutical development, and nuclear reactor design. He noted that “while Lawrence Livermore, Argonne, and Oak Ridge National Laboratories are three of the most powerful supercomputers in the world, and they are addressing incredibly important scientific issues that really require their advanced computing capabilities, Lawrence Berkeley’s National Energy Research Scientific Computing Center actually serves thousands of more users with only a fraction of those leadership machines’ computing power. The point is, not every computational research effort requires the two fastest, most sophisticated system we can possibly build, and I think we also need to do more to make what’s sometimes called ‘capacity’ supercomputing more accessible to both the academic and industrial research communities that could benefit.”

Four witnesses testified. Roscoe Giles, Chairman of the Advanced Scientific Computing Advisory Committee (ASCAC) in the DOE Office of Science summarized DOE’s Advance Scientific Computing research program and the work of the ASCAC. He provided recommendations from recent ASCAC reports. Also, he provided comments on the draft legislation stating that “it is essential that the effort to develop High End Computing be aggressively pursued and we expect that this legislation will be a key enabler of this.” He appreciated the outreach to industry and to the research community called for in the draft legislation as he promoted the Advanced Scientific Computing Research Computational Science Graduate Fellowship.

Rick Stevens, Associate Laboratory Director for Computing, Environment and Life Science at Argonne National Laboratory spoke about major challenges in exascale computing including significant increases in power consumption, increasing parallelism or the number of processing unites that can be applied to a single problem, the need for new programming models and increases in memory and reliability. “Exascale computing represents a critical technological and economic opportunity for the United States. Right now, the High Performance Computing global market is estimated at $10 billion, and that market is expected to grow to $40 billion over the next decade. At present, we lead the world both in the development of HPC and in the use of HPC for advancing science and engineering, and we are working hard to achieve the next great milestone.”

Donna Crawford, Associate Director for Computation at Lawrence Livermore National Laboratory described the role of exascale computing in national security and specifically the computational capability for support of the nuclear Stockpile Stewardship Program. She cautioned that “development of exascale-class systems cannot be achieved using straightforward refinement of today’s technology” as she emphasized the need for further research and development.

During his testimony, Daniel Reed, Vice President for Research and Economic Development at the University of Iowa provided recommendations for US investment in high-performance computing. Those recommendations emphasized a need for research and development to be allowed to develop without a pre-determined date for specific single system deployment and the need for research and development of both high-end data analytics and high-end computing. Other recommendations included the need to focus on next-generation algorithms and software applications and a need for collaboration between ASCR and the National Science Foundation, Department of Defense, National Institute of Standards and Technology and the National Institutes of Health.