Oak Ridge Leadership Computing Facility

Oak Ridge Leadership Computing Facility
Name	Oak Ridge Leadership Computing Facility
Established	2004
Location	Oak Ridge, Tennessee
Parent	Oak Ridge National Laboratory
Director	Thomas Zacharia
Website	Oak Ridge Leadership Computing Facility

Oak Ridge Leadership Computing Facility The Oak Ridge Leadership Computing Facility is a national user facility focused on petascale and exascale supercomputing resources for computational science. Located at Oak Ridge National Laboratory on the Oak Ridge, Tennessee campus, the facility supports large-scale simulation and data analysis across physical sciences, engineering, and computational biology. It provides access to leadership-class systems, user support, and development programs that connect researchers from universities, national laboratories, and industry.

History and development

The facility was established as part of initiatives by the U.S. Department of Energy and the Office of Science (United States Department of Energy) to accelerate high-performance computing, building on heritage from Enrico Fermi National Accelerator Laboratory collaborations and early programs at Argonne National Laboratory. Early milestones included awards managed through the Advanced Scientific Computing Research program and coordination with the Scientific Discovery through Advanced Computing (SciDAC) project. Leadership computing at Oak Ridge grew alongside national efforts exemplified by the National Strategic Computing Initiative and programs influenced by the High Performance Computing Act of 1991. Partnerships with industrial vendors such as IBM, Cray Inc., and NVIDIA shaped procurement and system architecture choices. Over time the facility aligned with exascale roadmaps articulated by the Exascale Computing Project and technical guidance from the National Laboratories consortium.

Facilities and architecture

The facility occupies space within the Oak Ridge National Laboratory campus and integrates with the Oak Ridge Leadership Computing Facility Data Center infrastructure, sharing power and cooling systems designed to support high-density compute racks. Its architecture decisions reflect trends from projects such as TOP500 and design influences from the Roadrunner (supercomputer) project and Blue Gene family systems. Networking is provisioned via high-performance fabrics interoperable with technologies developed by Mellanox Technologies and standards from the InfiniBand Trade Association. Storage systems incorporate resilience approaches developed at Los Alamos National Laboratory and file-system ideas tracing to Lawrence Livermore National Laboratory deployments. Security and access policies adhere to directives from the National Nuclear Security Administration for classified workflows and to the Office of Science for open scientific allocations.

Major supercomputers and systems

The facility has hosted a sequence of flagship systems, beginning with leadership systems influenced by designs from Cray Inc. and IBM such as iterations akin to Jaguar (supercomputer) and later systems resembling the Summit (supercomputer). These platforms combined technologies from NVIDIA GPU accelerators and CPU families like IBM POWER9 and processor roadmaps from Intel Corporation. System procurement and configuration drew on lessons from deployments at Argonne National Laboratory and Lawrence Berkeley National Laboratory. Benchmark results submitted to the TOP500 list and performance analyses by the High Performance Conjugate Gradients (HPCG) benchmark informed optimizations. Software stacks included MPI implementations from Open MPI and vendor MPI variants, compilers from GNU Compiler Collection and proprietary toolchains from Cray Inc. and IBM.

Research programs and applications

Research enabled by the facility spans simulations in areas championed by researchers at Massachusetts Institute of Technology, Stanford University, University of California, Berkeley, and Princeton University. Applications include climate modeling from groups associated with the National Oceanic and Atmospheric Administration, materials science studies linked to the Materials Genome Initiative, nuclear physics modeling cooperating with the Thomas Jefferson National Accelerator Facility, and fusion research supporting the Princeton Plasma Physics Laboratory. Computational chemistry and biology projects involved teams from Harvard University and Broad Institute, while energy systems modeling connected with Sandia National Laboratories and National Renewable Energy Laboratory. Data-intensive projects leveraged methods developed by collaborators at Carnegie Mellon University and University of Illinois Urbana-Champaign.

Collaborations and partnerships

The facility operates through partnerships with NVIDIA, Cray Inc., IBM, Intel Corporation, and Hewlett Packard Enterprise. It coordinates user programs with the Office of Science and joint initiatives with peer facilities at Argonne National Laboratory and Lawrence Livermore National Laboratory. International collaborations include exchanges with CERN, European Centre for Medium-Range Weather Forecasts, and university partners such as University of Cambridge and ETH Zurich. Workforce development and training programs have ties to Oak Ridge Associated Universities and education initiatives with University of Tennessee, Knoxville and Tennessee Tech University.

Operations, funding, and administration

Operational oversight is managed by Oak Ridge National Laboratory under contracts awarded by the U.S. Department of Energy with management arrangements influenced by UT-Battelle. Funding sources include allocations from the Office of Science, cooperative agreements with industry partners such as NVIDIA and Intel Corporation, and project grants from programs like the Exascale Computing Project. Administrative structures coordinate peer review and allocations through mechanisms similar to the Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program and scheduling policies used across National Laboratories. The facility supports user services, training, and code development with staff experienced in software engineering practices from institutions like Lawrence Berkeley National Laboratory and operational practices aligned with National Institute of Standards and Technology guidance.

Category:United States Department of Energy national laboratories