Optimizing Performance on Linux Clusters Using Advanced Communication Protocols: Achieving Over 10 Teraflops on a 8.6 Teraflops Linpack-Rated Linux Cluster
Conference
·
OSTI ID:965623
Advancements in high-performance networks (Quadrics, Infiniband or Myrinet) continue to improve the efficiency of modern clusters. However, the average application efficiency is as small fraction of the peak as the system’s efficiency. This paper describes techniques for optimizing application performance on Linux clusters using Remote Memory Access communication protocols. The effectiveness of these optimizations is presented in the context of an application kernel, dense matrix multiplication. The result was achieving over 10 teraflops on HP Linux cluster on which LINPACK performance is measured as 8.6 teraflops.
- Research Organization:
- Pacific Northwest National Laboratory (PNNL), Richland, WA (US)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- AC05-76RL01830
- OSTI ID:
- 965623
- Report Number(s):
- PNNL-SA-44417; KP1704020
- Country of Publication:
- United States
- Language:
- English
Similar Records
Building the World's Fastest Linux Cluster
Scalability and Performance of a Large Linux Cluster
PVFS : a parallel file system for linux clusters
Journal Article
·
Fri Oct 24 00:00:00 EDT 2003
· Published in: Clusterworld, vol. 1, no. 1, December 1, 2003, pp. 16-20,54
·
OSTI ID:15013813
Scalability and Performance of a Large Linux Cluster
Journal Article
·
Wed Jan 19 23:00:00 EST 2000
· Journal of Parallel and Distributed Computing
·
OSTI ID:750316
PVFS : a parallel file system for linux clusters
Conference
·
Thu Apr 27 00:00:00 EDT 2000
·
OSTI ID:754505