Yellowstone (supercomputer)

Yellowstone (supercomputer). It was a high-performance computing system dedicated to advancing research in the earth system sciences. Operated by the National Center for Atmospheric Research (NCAR) and funded by the National Science Foundation (NSF), it was installed at the Wyoming Supercomputing Center in Cheyenne, Wyoming. The system served as a primary resource for the U.S. academic community, enabling complex simulations of atmospheric, oceanic, and geological processes.

Overview

The deployment of this system represented a major investment by the NSF in computational infrastructure for environmental science. Managed by NCAR's Computational and Information Systems Laboratory (CISL), it was housed in a specially designed facility in Cheyenne, Wyoming. Its primary mission was to support the research needs of thousands of scientists across the United States, affiliated with institutions like the University Corporation for Atmospheric Research (UCAR). The system succeeded the earlier Bluefire system and was itself succeeded by the Cheyenne system.

Hardware and architecture

The system was built by IBM based on its iDataPlex server platform. It integrated a total of 72,288 processor cores from Intel's Xeon Sandy Bridge microarchitecture family. The interconnect utilized a Mellanox InfiniBand FDR network to facilitate high-speed communication between nodes. For storage, it employed an IBM General Parallel File System (GPFS) solution, with a massive DataDirect Networks (DDN) array providing over 16 petabytes of capacity. The entire installation required a sophisticated liquid cooling system to manage its substantial thermal output.

Computational capabilities

At its peak, the system achieved a theoretical performance of over 1.5 petaflops, as measured by the LINPACK benchmark. This computational power placed it prominently on the TOP500 list of the world's most powerful supercomputers upon its debut. Its architecture was particularly suited for large-scale climate model simulations, numerical weather prediction, and computational fluid dynamics studies. The extensive parallel computing environment allowed researchers to run ensemble forecasts and high-resolution global models that were previously infeasible.

Operational history and usage

The system entered full production service in late 2012 following an extensive acceptance testing period. Throughout its operational lifespan, it allocated millions of node-hours annually to projects through the NSF's XSEDE (Extreme Science and Engineering Discovery Environment) program. Major research campaigns included the CESM (Community Earth System Model) project and investigations led by the NOAA (National Oceanic and Atmospheric Administration). It was officially retired from service in January 2021, having been replaced by the more powerful Cheyenne system at the same facility.

Impact and scientific achievements

The system enabled transformative research across the geosciences. It was instrumental in producing high-resolution simulations for the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report. Scientists used it to study phenomena such as tropical cyclone intensification, Arctic sea ice loss, and regional impacts of climate change. Its computational resources also advanced the field of space weather prediction, supporting models developed by the NASA Community Coordinated Modeling Center. The vast datasets generated contributed to public archives like those held by the UCAR NCAR Research Data Archive.

Category:Supercomputers Category:National Center for Atmospheric Research Category:IBM supercomputers