A pragmatic approach of performance evaluation of MPI-OpenMP on a 12 nodes multi-core cluster
New problems related to scalability arise with the increase of number of Quad-Core-based clusters and the introduction of compute nodes that are designed with large memory capacity, shared by multiple cores. In this work, we perform experimental investigation of the overall performance of a cluster...
| Main Authors: | , |
|---|---|
| Format: | Book Chapter |
| Language: | English |
| Published: |
CRC Press, USA
2013
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/24253/ http://irep.iium.edu.my/24253/ http://irep.iium.edu.my/24253/1/Abdel_Sakib_NGNCRC_BookChapter.pdf |
| Summary: | New problems related to scalability arise with the increase of number of Quad-Core-based clusters and the introduction of compute nodes that are designed with large memory capacity, shared by multiple cores. In this work, we perform experimental investigation of the overall performance of a cluster built with nodes having a dual Quad-Core Processor on each one. Some benchmark results are presented and some observations are mentioned during handling such processors on a benchmark test. This is a reality that a Quad-Core-based cluster's complexity arises from the fact that both local communications and network communications between the running processes need to be addressed. Hence, the potentials of an MPI-OpenMP approach are pinpointed in our work because of its reduced communication overhead. At the end, we come to a conclusion that an MPI-OpenMP solution should be considered in such clusters since optimizing network communications between nodes is as important as optimizing local communications between processors in a multi-core cluster. This chapter is basically to show results from previous experiences on this issue which may be useful for future applications and cluster architects, which could be essential for building some crucial segments of the next-generation networks. |
|---|