Energy’s Supercomputing Program
Published: June 13, 2018
Energy expands its supercomputing capabilities with the unveiling of its newest supercomputer, Summit, paving the way for advances in energy research, artificial intelligence and scientific discovery.
The Department of Energy (DOE) has some of the greatest high performance computing (HPC) capabilities through the powerful supercomputers it develops and operates. In fact, Energy now has six of the ten fastest supercomputers in the world residing at its national labs. According to the federal agency’s computing site, “supercomputers are used to model and simulate complex, dynamic systems that would be too expensive, impractical or impossible to physically demonstrate.” Titan – the fourth fastest computer in the world and located at Oak Ridge National Laboratory (ORNL), has a peak performance exceeding 27,000 trillion calculations per second (27 petaflops) with a system memory of 710 terabytes.
Academia, industry and other federal agencies utilize DOE supercomputers for a variety of research in scientific and economic discovery. The Department of Veterans Affairs has partnered with Energy to use HPC resources to analyze a database with millions of veteran’s data to help prevent suicide and accelerate medical breakthroughs in cancer and cardiovascular disease. Likewise, Energy is devoting some of its supercomputing resources to early-stage research projects that aim to improve energy efficiency in transportation environments. The Big Data Solutions for Mobility program, for example, is working to develop tools and algorithms that use real-time data to model urban transportation systems for optimal energy use.
On June 8, 2018, Energy unveiled a new supercomputer at ORNL called Summit. Summit is eight times more powerful than Titan and now holds the title of the fastest computer in the world. The new system can perform 200,000 trillion calculations per second—or 200 petaflops at peak use and in certain science applications will have the ability for more than three billion-billion mixed precision calculations per second. According to agency’s announcement of the system, “Summit offers unparalleled opportunities for the integration of AI and scientific discovery, enabling researchers to apply techniques like machine learning and deep learning to problems in human health, high-energy physics, materials discovery and other areas.”
The Exascale Computing Project
The agency does not intend to stop there with its computing power. In collaboration with the Office of Science and National Nuclear Security Administration (NNSA), DOE’s Exascale Computing Project will work with national labs, other researchers and hardware vendors to deliver exascale computing systems. Referred to as Aurora or A21, the first exascale system is expected to go live at Argonne National Laboratory (ANL) by 2021 with the capability of one quintillion operations per second – or 1000-petaflop (1 exaflop). In April, the agency released a Request for Proposals (RFP) for two additional exascale systems worth up to $1.8B to be deployed at ORNL and Lawrence Livermore National Laboratory (LLNL) and delivered in the 2021-2023 time frame. The RFP lists the possibility of an additional system at ANL to be delivered by 2023, pending additional congressional funding. All new exascale systems are expected to be diverse from one another and provide 50 to 100 times greater performance than the current supercomputers.
According to an article on the exascale program, the computers are anticipated to perform many functions to excel areas of energy, science and national security, “exascale computers are expected to be able to design chemical catalysts with little to no experimentation and run core simulators that speed the development of new nuclear reactors. They could also help increase the productivity of wind farms… Exascale machines could also run national security applications, including modeling nuclear weapon performance based on stockpile stewardship practices.”
FY 2019 Budget Details
The FY 2019 request places emphasis on Energy’s exascale computing program with $636M allotted for exascale computing by 2021, a $376M increase above FY 2017 levels. Of that, $473M will go to the Office of Science to further develop the exascale computing system and continue site preparations for deployment in 2021 and 2022. The remaining $163M will go to the NNSA, including $24M for 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). Furthermore, $105M is requested under the Office of Science for quantum computing to build U.S. competency and competitiveness in quantum science, computing and sensor technology.