This is an old revision of this page, as edited by Runningonbrains (talk | contribs) at 21:55, 29 February 2012 (wl NCAR). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.
Revision as of 21:55, 29 February 2012 by Runningonbrains (talk | contribs) (wl NCAR)(diff) ← Previous revision | Latest revision (diff) | Newer revision → (diff)The Yellowstone is the name of the new petascale supercomputing resource in the NCAR-Wyoming Supercomputing Center (NWSC) in Cheyenne, Wyoming with production computing operations expected to begin in the summer of 2012.
Yellowstone will be an IBM iDataPlex supercomputer system, consisting of 4,662 nodes, each with two Intel Sandy Bridge EP processors, for a total of 74,592 processor cores and 149.2 terabytes of memory. Nodes will be connected using a Mellanox FDR InfiniBand full fat tree. Nodes will run Red Hat Enterprise Linux Version 6.
The central file and data storage resource, called GLADE, will consist of file system servers and storage devices that will be linked to the supercomputer systems and eventually provide 16.42 petabytes of usable disk space.
The Yellowstone system features 9.7 million times the computational rate, 3.4 million times the disk capacity, and 19 million times the central memory size of one of the world’s first supercomputers, the Cray 1-A, which supported NCAR’s computational science between 1977 and 1989.
References
- http://www.theregister.co.uk/2011/11/10/ncar_yellowstone_super_ibm
- http://www-03.ibm.com/press/us/en/pressrelease/35946.wss
- http://www.cesm.ucar.edu/management/SSC/Presentations/yellowstone.pdf