On February 12, 2018, the Trump administration submitted its Fiscal Year 2019 (FY-19) budget to Congress. The good news for the U.S. exascale program is that the numbers look very good and the support appears to be strong. There is also the interesting addition of a funding request to support research in quantum computing. One of the challenges with the FY-19 budget request is that it is being made before the FY-18 budget has been fully resolved. The good news is that it looks like all the pieces are in place to make that happen no later than March.
The proposed President’s FY-19 federal budget request identifies a total of $636M in the Department of Energy’s (DOE) budget for exascale activities. This is $376M above the FY-17 enacted level. Those funds are spread across two parts of the DOE, that in combination, make up the Exascale Computing Initiative (ECI). One element of the funding for exascale comes from the National Nuclear Security Adminstration (NNSA) Advanced Simulation and Computing (ASC) program with a request of $163M. The other part of the exascale computing request is for the Office of Science (SC) Advanced Scientific Computing Research (ASCR) program at $473M.
At this point, the only FY-19 budget documentation available on the DOE web site is the Budget Fact Sheet and the Budget in Brief, so details are limited. However, there are a few interesting facts. Starting at the high level, the NNSA ASC program request goes up by $40M to $703M from an annualized FY-18 Continuing Resolution (CR) of $659M. For the SC ASCR program, their requested funding goes up by a whopping $257M to a total of $899M over their annualized FY-18 CR funding of $642M. This number is impressive given that many other parts of the Office of Science saw significant reductions.
The introductory sections of the Budget in Brief document provides some interesting clues about the procurement and installation of the U.S. exascale computing systems. On page one, in the section labeled Accelerating Progress On National Priorities, there is a statement that the requested funds will be used at Argonne and Oak Ridge National Laboratories for R&D and design of exascale-capable systems. The section goes on to say that the first exascale system will be being deployed in 2021 and the second system, with a different architecture, deployed in 2022. This aligns with the previously announced Aurora system built by Intel at Argonne and the yet to be competed Frontier system at Oak Ridge. That procurement will use a similar process to CORAL and is expected to begin in early 2018,
Page two of the introduction provides more details about the NNSA element of the ECI. It states that of out of the $163M NNSA request for ECI that $24M goes towards the the Exascale Class Computer Cooling Equipment (EC3E) project at the Los Alamos National Laboratory (LANL), and $23M for the Exascale Computing Facility Modernization (ECFM) project at the Lawrence Livermore National Laboratory (LLNL). In the detailed NNSA budget breakdown, we learned that in FY-17 the first ASC exascale system passed its first mission needs Critical Decision point (CD-0) and is expected to be delivered in 2023. The system is code named “El Capitan” and will be installed at LLNL. This system will also be part of the follow-on CORAL procurement that will start in early 2018.
Another interesting statement in the Budget in Brief introduction is about DOE quantum computing activities. The statement is a paragraph about the $105M request to support “the emerging urgency of building U.S. competency and competitiveness in the developing area of quantum information science, including quantum computing and quantum sensor technology.” The Office of Science will administer these funds and both the ASCR program and the High Energy Physics (HEP) program budget descriptions mention something about quantum computing.
In summary – the FY-19 budget request looks very encouraging for the U.S. exascale program. The request is just the beginning of the budgetary process, but a strong start is always very important. Also, other developments in Congress with plans to resolve the FY-18 budget situation seem to indicate that Congress should be able to support (and perhaps improve upon) the President’s request numbers.
When we last left things with the FY-18 budget, the federal government and DOE was under a Continuing Resolution (CR) and they still are today. For ECI, that limited funding the lower of the levels provided by the House or Senate. For the ASC program, that was not a problem because those levels were $161M and were the same for both the House and Senate marks. The situation was a bit more complicated for the ASCR program because of their divergent marks. The House provided a total of $282M ($170M for ECP and $112M for facilities) and a total $434M in the Senate ($184M for ECP and $250M for facilities). The CR requires ASCR to spend at the lower House level, but to plan for and be ready to use the funds at the higher Senate level.
The passage of the House and Senate their individual appropriations bills towards the end of the fiscal year was encouraging that they could resolve their differences (known as a conference committee) and pass the bill, and thus complete the “regular order” process. Unfortunately, other issues outside of the budget process got in the way and that did not happen. As a result, the first CR was passed to run until December 8th, then the second one until December 22nd, then a third until January 19th, and a fourth that expired on February 8th. The federal government is currently under its fifth CR and that will expire on March 23rd (or almost six months after the start of the fiscal year). The good news is that the last CR included several important agreements that provide some hope that we may see actual signed FY-18 appropriation laws.
One of those agreements is to provide relief from the budget caps established by the Budget Control Act of 2011. Those limits (also known as sequestration) were applied (somewhat equally) to both the defense and discretionary elements of the federal budgets. The sequestration limits were causing major challenges for the Congressional appropriators as they distributed those limits on budget growth. With those limits eased, the appropriators will have a much easier time coming to an agreement. Another important element of the last CR was the decision to put off any discussion of the limit on the national debt until March of 2019. The debt limit puts a cap on how much more money the country can borrow the fund its operations. The U.S. was expected to reach its debt limit by early in 2018, which would have cause major problems with the FY-18 appropriations. Putting off that problem should make it easier to pass FY-18 appropriation. Finally, a third element of the latest CR was that it included the top-line numbers for the FY-19 appropriations. These are normally set in a Budget Resolution and are important to provide funding guidelines for the appropriations subcommittees.
Sorry for the budget “wonk” stuff, but the bottom line is that by March 23rd, the representatives of the 12 House and Senate appropriations subcommittees hopefully will have come together to resolve the differences between their bills. This will be much easier with the additional funding flexibility provided by the last CR. Once the differences are resolved, each of the appropriations bills will then have to be passed by the House and Senate and then set to the President for his signature.
Things continue to look good for the U.S. exascale program. The program has survived the transition of Presidential administrations and is getting strong support by the new Secretary of Energy and the Director of the Office of Management and Budget. The program has previously received support by the Senate and House appropriators, but there were challenges to balancing that support against other funding priorities. The flexibility provided by the latest CR should lead to a successful Senate and House conference on the appropriations bill and full Congressional passage and then Presidential signature. That will mean the return of “regular order” — although taking a bit more time than expected. As the saying goes . . . “Better late than never!”
About the Author
Alex Larzelere is a senior fellow at the U.S. Council on Competitiveness, the president of Larzelere & Associates Consulting and HPCwire’s policy editor. He is currently a technologist, speaker and author on a number of disruptive technologies that include: advanced modeling and simulation; high performance computing; artificial intelligence; the Internet of Things; and additive manufacturing. Alex’s career has included time in federal service (working closely with DOE national labs), private industry, and as founder of a small business. Throughout that time, he led programs that implemented the use of cutting edge advanced computing technologies to enable high resolution, multi-physics simulations of complex physical systems. Alex is the author of “Delivering Insight: The History of the Accelerated Strategic Computing Initiative (ASCI).”