Publication: Network Performance in Distributed HPC Clusters

All || By Area || By Year

Title	Network Performance in Distributed HPC Clusters
Authors/Editors*	Ben Huang, Michael Bauer, Michael Katchabaw
Where published*	The International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTAâ2005)
How published*	Proceedings
Year*	2005
Volume	-1
Number	-1
Pages	546-549
Publisher	CSREA Press
Keywords	High performance computing, grids, network performance analysis, benchmarking tools
Link
Abstract	Linux-based clusters have become prevalent as a foundation for High Performance Computing (HPC) systems. As these clusters become more affordable and available, and with the emergence of high speed networks, it is becoming more feasible to create HPC grids consisting of multiple clusters. One of the attractions of such grids is the potential to scale applications across the various clusters. This creates a number of interesting opportunities as well as introduces a number of challenges. A key question is the impact of the inter-cluster network on the overall performance of the computing environment. In this paper, we report on recent experiments to evaluate such performance.