What’s next for the National Strategic Computing Initiative? The HPC User Forum will start the process of answering that question by presenting the first public panel of speakers from NSCI lead agencies at its September meeting in Denver. Budgets, technology goals, industry collaboration, governance are all likely to be fair game at this meeting of heavy hitters from the HPC community drawn from academia, industry and government.
“We are getting a lot of interest from industry as they try to figure out exactly how they can connect into the effort,” said Bob Sorensen, a research vice president for HPC with IDC, who will moderate the panel. Created by a Presidential Executive Order on July 30, the NSCI is an expansive plan to accelerate and coordinate the US march towards exascale and advanced computing. Thus far, industry reaction has been extremely positive and eager for more details.
Announced panelists include: Randy Bryant, assistant director, Information Technology Research and Development, the Office of Science and Technology Policy; Irene Qualters, division director, Division of Advanced Cyberinfrastructure, National Science Foundation; Rob Leland, vice president, Science and Technology, and chief technology officer, Sandia National Laboratory; Rupak Biswas, deputy director of exploration technology at NASA Ames Research Center; Doug Kothe, director of the Consortium for Advanced Simulation of Light Water Reactors, Oak Ridge National Laboratory.
It’s likely there will also be panelists from the intelligence community said Sorensen.
“My main hope for this panel is to have an open, freewheeling, and thought-provoking forum to help establish key long-term relationships between US government HPC mission agencies and counterpart industrial and academic partners,” said Sorensen, noting the proposed national HPC strategy has two very ambitious and perhaps conflicting goals: ensuring that key US missions have access to the best high performance computers and helping build a robust commercial HPC sector.
Among the many questions Sorensen plans to ask panelists:
- What particular role does your agency plan on playing in this overall project?
- More specifically, which of the main objectives (exascale system, coherence between modeling and simulation with data analytics, post Moore’s Law technology development, enduring HPC ecosystem, and public private collaboration) do they see as their agency as presenting the greatest opportunity for their efforts. How will this translate into clearly defined activities and research efforts going forward?
- What will be the first steps taken to address these objective and what commitments of funding and other resources will be needed to ensure the success of these efforts?
- How will the respective agencies address the key goal of deploying and applying new HPC technologies broadly for economic competitiveness and scientific discovery? What steps will need to be taken by both government and industry partners to increase the overall prospects for economic impact, and what will be the most significant impediments to achieving that goal?
- How best can the agencies measure success in efforts as they move forward, i.e. what will be the clearest signs of early progress in the right direction?
Three agencies have been designated to lead NSCI: the Department of Energy (DOE), the Department of Defense (DOD), and the National Science Foundation (NSF). Also named are two foundational research and development agencies – the Intelligence Advanced Research Projects Activity (IARPA) and the National Institute of Standards and Technology (NIST). The five deployment agencies identified are the National Aeronautics and Space Administration, the Federal Bureau of Investigation, the National Institutes of Health, the Department of Homeland Security, and the National Oceanic and Atmospheric Administration.
An NSCI Executive Council has been charged with oversight and must produce an implementation plan by year-end. The HPC User Forum event is likely to be one of several outreaches to the HPC community as part of that effort. The NCSI Executive Council will be co-chaired by the director of the Office of Science and Technology Policy (OSTP) and the director of the Office of Management and Budget (OMB).
There are many critical steps to watch for to determine if this program is succeeding, said Sorensen. The level of funding is one. Fostering strong public-private research partnerships is another. Rationalizing the government procurement process to ensure a wide range of HPC options are explored and that risky or innovative technology developments are encouraged is also key.
“Such efforts will require close coordination across all procuring agencies to ensure that the most promising technologies are not overlooked and that there is a careful balance between meeting critical mission requirements and inculcating an entrepreneurial spirit into the procurement process. US government funding mechanisms and related procurement processes may need to be adjusted to encourage such behavior,” said Sorensen.
“For its part, industry must be realistic about which technologies have the best chance for widespread commercial success and, perhaps more important, be more willing to take a long-term perspective on R&D projects and their demonstrable results, financial or otherwise.”
Another pressing concern is developing sufficient HPC talent. “Too many HPC experts note that the HPC community is having trouble attracting new talent into this field, and many say that some sort of ‘Apollo Program’ could generate new and needed enthusiasm within the academic sector, creating a new cadre of HPC experts coming out of leading universities,” said Sorensen.
Arranging to host the panel is a coup for the HPC User Forum, which was established in 1999 to promote the health of the global HPC industry and address issues of common concern to users. The group has grown to 150 members, is directed by a volunteer steering committee of users from government, industry and academia, and operated for the users by market analyst firm IDC.
HPCwire will have comprehensive coverage of the panel.
At the September meeting, IDC will also present new findings from its ongoing study for DOE on HPC ROI. Other major topics in the two-day session include: energy research; weather/climate/environmental; high performance data analysis; and storage.