The Department of Energy is preparing to use its Frontier supercomputer to tackle 24 elementary science and engineering problems with applications and software suites.
The system, which is the cornerstone of the Agency’s work Exascale Computing Projectlate last month – albeit with some caveats – to regain the position in the name of The fastest exascale system in the world.
The Energy Department It now plans to scale it to about 9,400 nodes — discrete computers that make up a High Performance Computing (HPC) cluster, processing 1,880 petaflops — from the Oak Ridge Leadership Computing Facility Exascale computer To simulate problems of the national interest over the next eighteen months.
Exascale Computing Project (ECP) It is already running on about 200 nodes of border Hardware since January, with cabinets and a half of 74 exascale computers set aside for development teams to prepare and identify last-minute issues in the transition.
“It’s a tough time, and we’re very excited,” Doug Koth, director of ECP, told FedScoop. “Within the next couple of months, we will be launching Frontier and demonstrating very specific application and performance capabilities.”
Working with DOE sponsors, ECP has selected 24 primary science and engineering problems to focus on, even though future possibilities number in the hundreds or thousands, Koth said.
The problems span across major national pillars: economic, national and energy security; scientific discovery; And even health care – as long as it’s an exascal problem.
“The exascale problem is one that could not have been solved or dealt with without that kind of power,” Koth said. “It could mean that my concept-to-design cycle is a certain period, and being able to simulate some phenomena in design – with all the complexity I need to deliver a good solution – takes much longer than this design cycle.”
Engineers want results in days, hours, or minutes, so simulations that take months are not practical. Exascale computers can simulate more complex phenomena with higher confidence while allowing for much faster hypothesis cycles.
Energy production is a primary focus for the Department of Energy, so it will use ECP Frontier to simulate nuclear fusion and fission reactors, wind farms, power grids, the clean combustion of fossil fuels like coal, and internal combustion engines used by land and gas turbines. Power plants. ECP has a power grid implementation for power transmission and hopes to simulate a large portion of the country’s interconnections.
ECP will also use Frontier to answer basic science questions about the origin of the elements in the universe including astrophysics, neutron star mergers, and supernovae; the evolution of the universe known as cosmology; and fundamental forces of nature such as quantum chromodynamics.
The Stanford Linear Accelerated Laboratory has an ECP project that simulates how light sources interact with matter by shining photons from a free-electron laser through biological samples or metal alloys to understand their structure.
Unconventional applications include simulating the genome assembly of microbiomes, how materials respond in extreme conditions such as a radiation environment and docking scenarios with the COVID-19 virus.
“I am absolutely confident that we will see some major discoveries from this exascale era emerge with the applications we have developed,” Koth said.
ECP is part of the broader National Strategic Computing Initiative (NSCI) that the Department of Energy began in 2016, and its applications and software are not exascale—meaning they work on laptops, desktops, and collections, too.
Released over three years ago at E4S.io, the Extreme Scale Scientific Software Stack (E4S) consists of library applications that need to be innovated and sits on top of the Hewlett Packard Enterprise operating system.
The Department of Energy leads the NSCI, and its Industry Council and Agency target five agencies: the Department of Defense, the National Institutes of Health, the National Science Foundation, the National Oceanic and Atmospheric Administration, and NASA.
The Department of Education is working with the National Institutes of Health to develop machine learning for cancer treatment, called the Cancer Distributed Learning Environment (CanDLE), using Frontier.
Frontier applications may be built on multiple physical phenomena — fluid flow, heat transfer, and materials science — that can be exported to applications used by agencies outside the Big Five.
Frontier development was a token effort between ECP and the Oak Ridge Leadership Computing Facility, as is the Aurora and El Capitan exascale computers at Argonne and Lawrence Livmore National Laboratories, respectively. ECP establishes systems requirements.
“We understand very well what devices are being deployed,” Koth said. “And we design our apps to work well and take advantage of those devices.”