NERSC

NERSC
Name	National Energy Research Scientific Computing Center
Formation	1974
Type	Research center
Headquarters	Berkeley, California
Parent organization	Lawrence Berkeley National Laboratory
Region served	United States
Leader title	Director

NERSC

The National Energy Research Scientific Computing Center is a high-performance computing facility located at Lawrence Berkeley National Laboratory supporting large-scale scientific computing for a broad set of research programs. It provides computational resources, data storage, and user support to investigators associated with Department of Energy missions, national laboratories such as Argonne National Laboratory and Oak Ridge National Laboratory, and academic institutions including Massachusetts Institute of Technology and Stanford University. NERSC collaborates with initiatives and projects like Exascale Computing Project, Joint Genome Institute, Large Hadron Collider, and Human Genome Project-era efforts to enable simulation, modeling, and data analysis.

History

Founded in 1974, the center evolved alongside advances at institutions such as Lawrence Livermore National Laboratory and Los Alamos National Laboratory. Early systems paralleled the adoption of architectures from vendors like Cray Research and milestones in computational science highlighted by programs at National Aeronautics and Space Administration and National Institutes of Health. During the 1980s and 1990s, procurement cycles responded to trends exemplified by transitions at Sandia National Laboratories and projects including Accelerator Project developments. The center’s strategic shifts mirrored national initiatives such as the Advanced Scientific Computing Research program and collaborations with consortia like Oak Ridge Leadership Computing Facility.

Facilities and Infrastructure

The facility is housed within campus infrastructure at Lawrence Berkeley National Laboratory with data center specifications comparable to sites at Argonne National Laboratory and Fermilab. Power and cooling systems are engineered to standards used at Pacific Gas and Electric Company-adjacent campuses and adhere to requirements familiar to operators of National Renewable Energy Laboratory data centers. Networking connections utilize backbones interlinked with Energy Sciences Network and regional nodes at locations such as San Diego Supercomputer Center and University of California, San Diego. Storage architectures reflect approaches from peers like National Center for Supercomputing Applications and include tiered systems for projects related to CompleX Systems Initiative and large experimental collaborations like Laser Interferometer Gravitational-Wave Observatory.

Systems and Performance

Hardware procurements have included systems from vendors affiliated with programs at Cray Inc., IBM, and companies that supplied bolt-on accelerators used in efforts at NVIDIA-accelerated centers. Performance metrics are benchmarked with suites comparable to those used at TOP500 sites and reported in contexts similar to Green500 and SPEC evaluations. Allocated cycles support workloads ranging from climate modeling in partnership with National Oceanic and Atmospheric Administration researchers to materials simulation efforts resonant with projects at Argonne National Laboratory’s materials science programs. Peak computational throughput and sustained performance figures are routinely compared against leadership systems at Oak Ridge National Laboratory and European centers like Jülich Research Centre.

Research and User Community

The user base includes researchers from institutions such as University of California, Berkeley, Princeton University, University of Michigan, California Institute of Technology, and national laboratories including Brookhaven National Laboratory and Pacific Northwest National Laboratory. Scientific domains supported encompass astrophysics collaborations tied to Sloan Digital Sky Survey, fusion energy projects linked with Princeton Plasma Physics Laboratory, and genomics work associated with Broad Institute. Training and outreach echo programs at Computational Science Graduate Program and workshops coordinated with societies like American Physical Society and Association for Computing Machinery. User allocations and campaigns have enabled discoveries referenced alongside publications in journals like Nature, Science, and Physical Review Letters.

Software and Services

The center provides software stacks and user services that include batch scheduling and resource management comparable to systems running Slurm Workload Manager and middleware used in federations with Globus data transfer tools. Development environments support compilers and libraries familiar to users of MPI-based codes, productivity tools similar to those in GitHub repositories, and domain-specific frameworks used by teams at CERN and NASA Ames Research Center. Data management services parallel offerings at National Energy Technology Laboratory and incorporate provenance, metadata, and workflow systems akin to those developed by Kitware and academic collaborations with University of Illinois Urbana-Champaign.

Governance and Funding

Operational oversight is provided by leadership drawn from Lawrence Berkeley National Laboratory management and advisory boards that include representatives affiliated with U.S. Department of Energy program offices and scientific stakeholders from institutions like Massachusetts Institute of Technology and University of California campuses. Funding is primarily from programs within the Office of Science (United States Department of Energy) and coordinated with investments under initiatives such as the Advanced Scientific Computing Research portfolio and cross-laboratory partnerships that mirror funding structures at National Nuclear Security Administration-supported centers. Resource allocation and strategic planning involve community proposals evaluated under peer review processes analogous to review panels at National Science Foundation and programmatic guidance from entities such as White House Office of Science and Technology Policy.

Category:Supercomputer sites