In high-performance computing, what are Rmax and Rpeak?
In high-performance computing, Rmax and Rpeak are scores used to rank supercomputers based on their performance using the LINPACK Benchmark. A system's Rmax score describes its maximal achieved performance; the Rpeak score describes its theoretical peak performance. Values for both scores are usually represented in teraFLOPS or petaFLOPS.
According to TOP500.org (as of June 2013):
Titan, a Cray XK7
system operated by Oak Ridge National Laboratory (ORNL), achieved the highest Rmax
(17.59 petaFLOPS) and Rpeak (20.13 petaFLOPS) scores in the US.
- Big Red II, Indiana University's Cray XK7
system, achieved the second highest Rmax (597.4 teraFLOPS) and Rpeak
(1 petaFLOPS) scores among university-owned supercomputers. Conte, a
Hewlett-Packard Cluster Platform system run by Purdue University Research
Computing, posted the highest scores among university-owned
computers: 961.9 teraFLOPS (Rmax), 1.34 petaFLOPS (Rpeak).
- The highest scores achieved by digital services on the Extreme
Science and Engineering Discovery Environment (XSEDE) were:
XSEDE system Rmax Rpeak Stampede (TACC) 5.17 petaFLOPS 8.52 petaFLOPS Kraken (NICS) 961.9 teraFLOPS 1.17 petaFLOPS Keeneland (Georgia Tech) 319.6 teraFLOPS 614.5 teraFLOPS Gordon (SDSC) 285.8 teraFLOPS 336.1 teraFLOPS Lonestar (TACC) 251.8 teraFLOPS 301.8 teraFLOPS
For more about XSEDE compute, advanced visualization, storage, and special purpose systems, see the Resources Overview, Systems Monitor, and User Guides. For scheduled maintenance windows, outages, and other announcements related to XSEDE digital services, see User News.
This document was developed with support from National Science Foundation (NSF) grant OCI-1053575. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NSF.
This document is based upon work supported in part by National Science Foundation (NSF) grant 0910812 to Indiana University for "FutureGrid: An Experimental, High-Performance Grid Test-bed." FutureGrid project partners include the University of California - San Diego and the San Diego Supercomputer Center (SDSC), the University of Chicago/Argonne National Labs, the University of Florida, Purdue University, the University of Southern California, the University of Texas - Austin, and the Center for Information Services and High Performance Computing at Technische Universität Dresden.
Last modified on June 27, 2013.