UET scheduling with unit interprocessor communication delays
Discrete Applied Mathematics
A bridging model for parallel computation
Communications of the ACM
Analyzing the behavior and performance of parallel programs
Analyzing the behavior and performance of parallel programs
LogP: a practical model of parallel computation
Communications of the ACM
Network performance modeling for PVM clusters
Supercomputing '96 Proceedings of the 1996 ACM/IEEE conference on Supercomputing
LoGPC: Modeling Network Contention in Message-Passing Programs
IEEE Transactions on Parallel and Distributed Systems
Network performance-aware collective communication for clustered wide-area systems
Parallel Computing - Clusters and computational grids for scientific computing
Fast Measurement of LogP Parameters for Message Passing Platforms
IPDPS '00 Proceedings of the 15 IPDPS 2000 Workshops on Parallel and Distributed Processing
DiP: A Parallel Program Development Environment
Euro-Par '96 Proceedings of the Second International Euro-Par Conference on Parallel Processing-Volume II
Contention-Aware Communication Schedule for High-Speed Communication
Cluster Computing
Modeling advanced collective communication algorithms on cell-based systems
Proceedings of the 15th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming
| Hi-index | 0.00 |
One of the most important collective communication patterns for scientific applications is the many to many, also called complete exchange. Although efficient All-to-All algorithms have been studied for specific networks, general solutions like those found in well known MPI distributions are strongly influenced by the congestion of network resources. In this paper we present our approach to model the performance of the All-to-All collective operation. Our approach consists in identifying a contention factor that characterises the network environment, and using it to augment a contention-free communication model. This approach allows an accurate prediction of the performance of the All-to-All operation over different network environments with a small cost. Indeed, we demonstrate the accuracy of our approach by presenting our experiments with three different network environments, Fast Ethernet, Giga Ethernet and Myrinet.