posted Monday, February 19, 2007 7:08 PM by dongarra | 0 Comments

  • High Performance Linpack (HPL)

  The Top500 is a list computers ranked by their performance on the HPL Benchmark.
  HPL is a software package that solves a (random) dense linear system in double
  precision (64 bits) arithmetic on distributed-memory computers. It can thus be
  regarded as a portable as well as freely available implementation of the
  High Performance Linpack Benchmark.

Current Status

 Currently we have compiled HPL with both Visual Studio 2005 and Subsystem for
 UNIX Applications (SUA).  We have run HPL successfully on all of our interfaces
 which include: GigE, Mellanox's Infiniband, Silverstorm's Infiniband, and Myricom's Myrinet. 
 We have also compiled HPL against both Intel's MKL and AMD's ACML BLAS libraries to
 compare performance results since our cluster runs AMD Opteron processors.
 The ACML BLAS performs much faster.  The only thing we have been unsuccessful at
 running is HPL with Winsock Direct (WSD). 
 Currently achieving 242.4 GFlops on 96 processors.

Future Work

 Upcoming work on HPL will focus primarily on understanding performance differences
 between Windows CCS and Linux. Based on the findings we will tune the code.
 We're interested in creating tuning guidelines for various interconnects and the
 BLAS libraries in use. Also, we will attempt to get HPL running with WSD so that
 we have a better comparison of results against Linux.

For more information about the Top500 and HPL,
visit the respective websites at
http://www.top500.org/
and
http://www.netlib.org/benchmark/hpl/