Optimization of Communication in MPI-Based Clusters
Rashid Hassani, Ganesh Chavan, Peter Luksch · 2014
Many parallel applications in high performance computing have to communicate via wide area network, e.g., in a grid or Cloud environment that spans multiple sites. Communication of links across wide area network slows down the application due to high latency and low bandwidth. Much of this overhead is due to the current implementations of the message passing interface standard. Efficient operation of such a distributed system requires that the characteristics of communication and synchronization across wide area network connections be addressed appropriately. This paper proposes a new approach to refine Open MPI communication scheme by using an optimized algorithm to improve the communication performance of high performance computing applications within the message passing interface protocol stack in wide area network environment. Our technique reduces unnecessary communications and synchronizations between processors with optimized communication-computation overlap. Our results emphasize the validity and effectiveness of our technique in which significant improvement over traditional Open MPI has been achieved.