OLCF

While ORNL's Oak Ridge Leadership Computing Facility often makes headlines for hosting the world's fastest supercomputers, its ultimate achievement is more complex: helping researchers solve some of the world’s most challenging scientific problems.

By offering scientists both world-class high-performance computing and world-leading expertise in scientific computing, the OLCF is a rare asset among DOE's array of institutions — with only the Argonne Leadership Computing Facility offering similar services.

The OLCF's primary systems are the 200-petaflop Summit — able to perform 200 million billion calculations every second — and the exascale-class Frontier — which will soon be available to users and is capable of more than a billion billion calculations each second. They attract scientists from around the world for open science projects that tackle the most computationally complex problems, from diverse scientific domains such as materials sciences, Earth sciences, astrophysics and biology. And with its quantum computing program, the OLCF also offers researchers a sneak peek into the future of scientific discovery.

Summit

Since its launch in 2018, the Summit supercomputer has been a world leader in not only computational power but also in scientific achievement. To date, the IBM AC922 system has served up over 113 million hours of computing time. Even with the advent of  faster computers such as Frontier, Summit should prove invaluable for years to come.

Most users for OLCF systems are selected through highly competitive programs administered by DOE, primarily the annual Innovative and Novel Computational Impact on Theory and Experiment, or INCITE, program, run by DOE’s Office of Science. These awards pursue transformational advances in science and engineering and account for 60 percent of the available time on the leadership-class supercomputers at the OLCF and ALCF.

“INCITE allocations really serve as a bellwether for the next frontiers in advanced computing. The current class features a diverse portfolio of ambitious research campaigns representing the most advanced techniques in HPC in support of a broad range of both applied and basic research,” said Gina Tourassi, director of the National Center for Computational Sciences, which houses the OLCF. “We are proud to provide full-scale access to the world’s most powerful systems to our users at the leading edge in their science domains.”

Meanwhile, DOE's Office of Advanced Scientific Computing Research conducts the annual ASCR Leadership Computing Challenge. It grants one-year allotments of computing time to scientists from industry, academia and national laboratories whose work advances scientific and technological research in DOE mission areas such as fusion energy, geosciences, high energy physics and materials sciences.

Among its many success stories is a groundbreaking project led by University of Michigan aerospace engineering professor Venkat Raman to develop a rotating detonation engine, which detonates its fuel to create a wave front that continuously rotates within a cylindrical combustor. Many aerospace companies and governments around the world are pursuing this technology as well, but access to Summit has proved to be a critical advantage for his project, Raman said.

“To me, Summit is a step change. Its compute power is something I did not expect. Even the biggest calculations that we thought we were going to run would fit on 20 nodes of Summit — and Summit is over 4,000 nodes. So we really had to increase our ambition once we got access to the machine,” Raman said. “I think that sort of step change requires you to rethink how you do your simulations and what simulations can actually be done. We changed our algorithmic frameworks, we changed our codes, we even changed how we answer the questions.”

Quantum computing

Administered by the OLCF, the Quantum Computing User Program, or QCUP, awards time on quantum computers owned by companies such as IBM, Quantinuum (formerly Honeywell) and Rigetti Computing to scientists with quantum-specific projects. Although the systems are not in-house, OLCF administrators benefit by getting first-hand experience at managing open science projects on cutting-edge quantum computers.

“As a facility, we’re trying to understand when quantum computing as a technology will be ready for prime time," said ORNL’s Travis Humble, manager of the user program and interim director of the Quantum Science Center. "At the moment, it’s very experimental, with a focus on exploration and discovery. We established the user program to monitor the technology and track the progress that the users are making."

The 4-year-old program has supported over 100 users running more than 70 projects by providing compute time from a variety of companies that are producing quantum computers.

“One of the things that differentiates QCUP is that we’re not advocating for a particular quantum computing system. We’re very agnostic to the individual devices,” Humble said. “We really just want people to do the best science.”

Fengqi You, the Roxanne E. and Michael J. Zak Professor in Energy Systems Engineering in Cornell University’s College of Engineering, used the program to successfully test a proposed quantum computer-based artificial intelligence system for identifying and diagnosing faults in electrical power grids. Being able to access an actual quantum computer from D-Wave Systems through QCUP has been a big boost to designing his new detection system.

“Quantum computers can help improve certain aspects of the training methodology by providing better gradient estimates compared to their classical counterparts, thus improving convergence and resulting in better generalization,” You said. “Apart from the improvement in training techniques, the proposed fault diagnosis framework also demonstrates faster response times than popular fault detection methods.”

Frontier

After years of planning and construction, the world’s most powerful supercomputer — Frontier — brings an eightfold increase in computational power over Summit, allowing researchers to tackle more extensive problems and answer more complex questions than ever before.

Early access to Frontier and its test platforms were granted through the OLCF's Center for Accelerated Application Readiness program, or CAAR. Through the program, the OLCF partners with application core developers, vendor partners, and OLCF staff members to optimize scientific applications for exascale performance, ensuring that Frontier will perform large-scale science on day one of full user operations. To be selected for the CAAR program, projects had to show a high potential for scientific advancement that could be achieved on petascale computers like Summit.

Eight CAAR projects were chosen for Frontier, with simulations ranging in scale from protons to our galaxy.

“Exascale computing gives us tremendous opportunities to make scientific advancements in plasma physics that were utterly beyond our reach before,” said Sunita Chandrasekaran, an associate professor at the University of Delaware who is developing an open-source simulations framework for plasma and laser-plasma physics with applications in radiation therapy, high-energy physics, and photon science. “Also, it allows us to do more simulations. So instead of doing a single simulation once, we can play with the problem and find out which physics conditions work best. For tumor therapy this means we can discover new, innovative ways to improve the energy, quality and precision of proton beams.”

Evan Schneider, an assistant professor at the University of Pittsburgh, has been developing her Cholla astrophysics simulation code on a variety of different systems over the years — but Frontier is a "galaxy-changer," offering unprecedented resolution.

"Using our Cholla astrophysics software on an exascale machine, we will be able to run a simulation where we can directly resolve what's happening in small patches of the galaxy,” Schneider said. “So instead of having to make assumptions about how the stars are forming and how the supernovae are affecting the galaxy, we can actually just directly simulate that using the physics that we understand. There's no other computer that's big enough and has enough computing power to get the resolution that we require.”