Blue Waters (supercomputer)
This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| Blue Waters (supercomputer) | |
|---|---|
.png) | |
| Name | Blue Waters |
| Manufacturer | Cray Inc. |
| Developer | National Center for Supercomputing Applications |
| Released | 2012 |
| Discontinued | 2019 |
| Power | 3.9 MW |
| Os | CNK, Linux |
| Memory | 1.5 PB (aggregate) |
| Storage | 26 PB Lustre |
| Speed | 13.34 PFLOPS (theoretical) |
| Type | Petascale supercomputer |
Blue Waters (supercomputer) Blue Waters was a petascale supercomputer deployed at the National Center for Supercomputing Applications in Urbana–Champaign. It delivered large-scale computational resources to researchers from institutions such as the University of Illinois Urbana-Champaign, Argonne National Laboratory, Lawrence Livermore National Laboratory, Los Alamos National Laboratory, and users funded by agencies including the National Science Foundation, Department of Energy, and NASA. Designed by Cray Inc. and managed with participation from entities like the State of Illinois and the University of Illinois Foundation, Blue Waters supported simulations, modeling, and data analysis across domains spanning astrophysics, climate science, and bioinformatics.
Overview
Blue Waters provided capability-class computing, aiming to support sustained petascale calculations for Grand Challenge projects from organizations such as the National Aeronautics and Space Administration, National Oceanic and Atmospheric Administration, Department of Defense, National Institutes of Health, and international collaborators like CERN and the European Space Agency. The system combined high-performance compute nodes, large-scale storage arrays, and a high-speed network fabric to enable workflows from research groups at institutions including MIT, Stanford University, Princeton University, Harvard University, and Caltech. Blue Waters facilitated collaborations involving principal investigators and research teams who had previously used supercomputers at centers like Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, and Pacific Northwest National Laboratory.
Architecture and Hardware
Blue Waters' architecture integrated commodity and custom elements with contributions from vendors including Cray Inc., NVIDIA, Intel Corporation, AMD, and storage partners such as Seagate Technology. The machine featured hybrid compute cabinets composed of multi-core x86 processors from Intel Xeon families and accelerator technologies related to NVIDIA Tesla GPUs and prototypes comparable to later FPGAs deployments. Interconnect topology leveraged high-bandwidth fabrics akin to technologies from Infiniband Trade Association implementations and Cray's proprietary interconnect lineage from systems like the Cray XT5 and Cray XE6. Cooling, power distribution, and facility design were influenced by standards from groups such as the American Society of Mechanical Engineers and the U.S. Department of Energy facility best practices. Storage subsystems used parallel file system concepts similar to Lustre (file system) and metadata management techniques researched at institutions such as Los Alamos National Laboratory and Argonne National Laboratory.
Software and Programming Environment
The software stack for Blue Waters included system software and user-level tools developed in collaboration with projects from OpenMP Architecture Review Board, the Message Passing Interface Forum, and software ecosystems represented by organizations like Khronos Group and Apache Software Foundation. Compilers and toolchains from GNU Project, Intel Corporation, and performance libraries similar to BLAS and LAPACK were part of the environment, while debuggers and profilers reflected capabilities seen in TotalView and Valgrind tool suites. Workflow and job management integrated schedulers akin to PBS Professional and resource managers inspired by lessons from SLURM Workload Manager. Scientific packages and community codes ported to Blue Waters came from projects such as GROMACS, LAMMPS, FLASH (software), Enzo (software), NAMD, WRF (Weather Research and Forecasting Model), and visualization tools comparable to ParaView and VisIt.
Performance and Benchmarks
Blue Waters achieved theoretical peak performance in the petaflop range and sustained performance reported for real-world applications approached multi-petaflop scales in parallel workloads. Benchmarking activities referenced standards and suites like LINPACK, HPCC, and domain-specific validations akin to those used by teams at NASA Ames Research Center, European Centre for Medium-Range Weather Forecasts, and National Center for Atmospheric Research. Performance tuning drew on methods developed by researchers affiliated with Sandia National Laboratories, Argonne National Laboratory, and university groups at University of California, Berkeley and University of Illinois at Urbana-Champaign to optimize memory hierarchy, I/O patterns, and interconnect exploitation.
Development and Funding
Blue Waters resulted from a multi-year proposal process involving agencies such as the National Science Foundation, with programmatic oversight from NSF divisions and advisory panels including members from National Research Council committees and scientific advisory boards drawn from institutions like Johns Hopkins University, Columbia University, and University of Chicago. Funding combined federal support through awards resembling NSF Major Research Instrumentation grants with state contributions from the State of Illinois and infrastructure investment by the University of Illinois. Industry partnerships and procurement involved contracting processes similar to those used by General Services Administration for federal acquisitions, and collaborations with vendors including Cray Inc., Dell Technologies, and storage firms akin to EMC Corporation.
Operational History and Decommissioning
Blue Waters entered full production in the 2010s and supported thousands of projects led by investigators at universities such as Yale University, University of Michigan, Columbia University, and Purdue University. Operational management drew on systems engineering practices from Carnegie Mellon University research and IT governance frameworks used by institutions like Indiana University. The system was decommissioned after an operational lifecycle influenced by evolving technology roadmaps and successor initiatives at centers such as Ames Research Center and national programs sponsored by the Department of Energy and National Science Foundation. Decommissioning processes mirrored procedures followed at other facilities like Oak Ridge National Laboratory when retiring capability-class machines.
Impact and Notable Applications
Blue Waters enabled breakthroughs in computational astrophysics linked to researchers at Space Telescope Science Institute and projects exploring cosmological structure formation alongside teams from University of California, Santa Cruz. Climate and geoscience simulations conducted on Blue Waters informed studies associated with Intergovernmental Panel on Climate Change authors and researchers from NOAA Geophysical Fluid Dynamics Laboratory. In biology and chemistry, molecular dynamics studies related to National Institutes of Health-funded efforts and pharmaceutical collaborations used codes similar to CHARMM and AMBER. The system's legacy influenced subsequent procurements and architectures at centers like Argonne Leadership Computing Facility and contributed to workforce training initiatives at universities including Georgia Institute of Technology and University of Texas at Austin.