Skip to main content

The Oak Ridge Leadership Computing Facility at Oak Ridge National Laboratory (ORNL) is home to America’s first exascale system, the 2 exaflops HPE Cray EX Frontier supercomputer. Exascale is the next level of computing performance. By solving calculations more than five times faster than today’s top supercomputers—exceeding a quintillion, or a billion billion, calculations per second—exascale systems will enable scientists to develop critically needed technologies for energy, medicine, materials, and more.

ORNL has decades of experience in the delivery and operation of world-leading supercomputers for scientific discovery. Since 2006, the Oak Ridge Leadership Computing Facility has deployed four supercomputers that debuted as fastest in the world, including Summit in 2018. In May 2022, Frontier came online as the first exascale machine in the world. Its power will help researchers answer problems of national importance that cannot be addressed on existing supercomputing platforms, such as:

  • enhance nuclear reactor efficiency and safety by modeling their entire life span;
  • uncover the underlying genetics of disease;
  • discover patterns in patient data for precision medicine; and
  • further integrate artificial intelligence with data analytics, modeling, and simulation.

Frontier System Support

The System User Guide is the definitive source of information about Frontier, and details everything from connecting to running complex workflows. Please direct questions about Frontier and its usage to the OLCF User Assistance Center by emailing [email protected].

System Specifications

Compute Node:1 64-core AMD “Optimized 3rd Gen EPYC” CPU 4 AMD Instinct MI250X GPUs
GPU Architecture:AMD Instinct MI250X GPUs, each feature 2 Graphics Compute Dies (GCDs) for a total of 8 GCDs per node
System Interconnect: 4-port HPE Slingshot 200 Gbps (25 GB/s) NICs providing a node-injection bandwidth of 800 Gbps (100 GB/s)
High-Performance Storage: 700 PB HDD+11 PB Flash Performance Tier, 9.4 TB/s and 10 PB Metadata Flash Lustre
Programming Models: MPI, OpenMP, OpenACC, CUDA, OpenCL, DPC++, HIP, RAJA, Kokkos, and others
Node Performance: 53 TF double precision
System Size: 9,402 nodes

CAAR Projects

In preparation for the Frontier supercomputer, OLCF has selected eight research projects to participate in the Frontier Center for Accelerated Application Readiness (CAAR) program. Through Frontier CAAR, the OLCF are partnering with application core developers, vendor partners, and OLCF staff members to optimize simulation, data-intensive, and machine learning scientific applications for exascale performance, ensuring that Frontier will be able to perform large-scale science when it opens to users in 2022. Consisting of application core developers and staff from the OLCF, the partnership teams are receiving technical support from Cray and AMD—Frontier’s primary vendors—and are accessing multiple early-generation hardware platforms prior to system deployment.

The Frontier CAAR teams include:

Cholla simulation

Code: Cholla

PI: Evan Schneider, University of Pittsburgh

Description: Massive amounts of new data about the Milky Way through surveys such as GALFA HI and GAIA mean that a more powerful computing system like Frontier will be necessary to reproduce all the data in a simulation, as well as make detailed astrophysical predictions. Schneider’s team is working to simulate a Milky Way-like galaxy using Cholla, an astrophysical simulation code developed for the extreme environments encountered in astrophysical systems.

CoMet data visualization

Code: CoMet (Combinatorial Metrics)

PI: Daniel Jacobson, Oak Ridge National Laboratory

Description: Using the CoMet application, researchers can conduct, among other things, large scale Genome-Wide Epistasis Studies (GWES), co-evolution and pleiotropy studies. Current work with CoMet is being used in projects ranging from climatype clustering and bioenergy to clinical genomics, including the study of the genetics of opioid addiction and toxicity, chronic pain, Alzheimer’s, and autism. With Frontier, Jacobson’s team will be able to solve scientific problems within much more complex biological systems than was previously possible on Summit. In addition, the ability to perform climatype comparisons for every year over a decades of time rather than a single comparison based on the mean of 30 years of data will allow researchers to gain a better understanding of the biological systems related to bioenergy, agriculture, and medicine.

Code: GESTS (GPUs for Extreme-Scale Turbulence Simulations)

PI: P. K. Yeung, Georgia Institute of Technology

Description: Understanding fluid turbulence and the potential universal properties of turbulence is a grand challenge in many fields of science and engineering, from pollutant dispersion and combustion to ocean dynamics and astrophysics. Using the GESTS code, P. K. Yeung’s team performs direct numerical simulations of turbulent fluid flows by computing instantaneous velocity fields in time and 3D space with a wide range of scales according to the Navier-Stokes equations, which express the exact physical principles of conservation of mass and momentum. On Frontier, the team will be able to produce a simulation of forced isotropic turbulence with nearly 35 trillion grid points, compared to 6 trillion they’ve achieved on Summit.

Code: LBPM (Lattice Boltzmann Methods for Porous Media)

PI: James Edward McClure, Virginia Polytechnic Institute and State University

Description: To study the behavior of physical processes within rock structures, specifically to model multiphase flow processes, James McClure and his team have to first understand heterogeneous wettability. Wettability refers to the surface energy between fluids and solids and can vary due to the roughness and mineral composition of the material. McClure’s team can use LBPM to read the volumetric maps of mineral composition and assign local wetting properties accordingly. With Frontier, they hope to train neural networks to predict the future geometric configuration of fluids.

Code: LSMS (Locally-Selfconsistent Multiple Scattering)

PI: Markus Eisenbach, Oak Ridge National Laboratory

Description: To understand states that go beyond a periodic crystalline lattice structure, researchers like Markus Eisenbach must be able to simulate the many thousands of atoms needed to describe extended electronic and magnetic orderings. In previous research it has been necessary to employ classical models due to the scaling issues that arise with first principles (FP) calculations. With LSMS—a code for FP calculations of alloys and magnetic systems—in conjunction with the computational power of Frontier, however, Eisenbach’s team can run previously inaccessible calculations of realistic condensed matter systems from FP.

model of protocell

Code: NAMD (Nanoscale Molecular Dynamics )

PI: Emad Tajkhorshid, University of Illinois at Urbana-Champaign

Description: Using Frontier in conjunction with NAMD, a molecular dynamics code designed for the simulation of large biomolecular systems, Emad Tajkhorshid and his team can study the entry process of the Zika Virus (ZIKV). A greater understanding of the way viruses like ZIKV enter host cells through clathrin-dependent endocytosis may pave the way for new drugs and vaccines to prevent future outbreaks. Frontier’s enormous computational power means that these simulations can be run with a great degree of accuracy across multiple models.

Code: NuCCOR (Nuclear Coupled-Cluster Oak Ridge)

PI: Morten Hjorth-Jensen, Michigan State University

Description: NuCCOR is a nuclear physics application that uses coupled cluster techniques to describe many-body systems, or microscopic systems consisting of a large number of interacting particles. Using Frontier, Hjorth-Jensen’s team will no longer be limited to performing computations only on atomic nuclei that exhibit subshell closures for neutrons and protons, and instead can take an alternative approach using symmetry-projection techniques. They will now be able to study complex time-dependent phenomena such as nuclear reactions and fission.

Code: PIConGPU (Particle-in-cell on Graphics Processing Units)

PI: Sunita Chandrasekaran, University of Delaware

Description: PIConGPU is an Open Source simulations framework for plasma and laser-plasma physics used to develop advanced particle accelerators for radiation therapy of cancer, high energy physics and photon science. With Frontier, Chandrasekaran’s team will significantly improve both on the predictive capabilities of simulations and on reaching particle energies suitable for application. The complex plasma dynamics that govern the final particle beam properties can now be studied at an unprecedented temporal and spatial resolution. This will provide accelerator development with the high quality data needed to advance plasma accelerators towards application.

Frontier

The Frontier supercomputer at the Department of Energy’s Oak Ridge National Laboratory earned the top ranking today as the world’s fastest on the 59th TOP500 list, with 1.1 exaflops of performance. The system is the first to achieve an unprecedented level of computing performance known as exascale, a threshold of a quintillion calculations per second.

Latest Frontier Highlights

Filter

Uniting to Build Solutions

Two ORNL-led computing teams were recognized with Department of Energy Secretary’s Honor Awards, presented by Secretary Jennifer Granholm on January 8, 2025, at the DOE Forrestal building. Secretary of Energy Jennifer Granholm delivered the Secretary’s…
Katie BetheaKatie BetheaJanuary 17, 20254 min
winners of the 2024 Gordon Bell Prize are awarded on stage

Game-Changing Quantum Chemistry Calculations on Frontier Earn Gordon Bell Prize

This year’s Association for Computing Machinery’s Gordon Bell Prize in supercomputing goes to researchers led by the University of Melbourne who used the Frontier supercomputer to conduct a quantum molecular dynamics simulation 1,000 times greater…
Jeremy RumseyJeremy RumseyNovember 22, 20244 min
Award recipients of the 2024 Gordon Bell Prize for Climate Modelling stand on stage

Gordon Bell Climate Prize Goes to KAUST Frontier Users’ Exascale Climate Emulator

The 2024 Gordon Bell Prize for Climate Modelling has been awarded to a team of researchers led by the King Abdullah University of Science and Technology, or KAUST, Saudi Arabia, who used the Frontier supercomputer…
Jeremy RumseyJeremy RumseyNovember 22, 20245 min