The documents distributed by this server have been provided by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a noncommercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.
Publications of SPCL
|T. Hoefler and M. Snir:|
|Generic Topology Mapping Strategies for Large-scale Parallel Architectures|
(. Vol , Nr. , In Proceedings of the 2011 ACM International Conference on Supercomputing (ICS'11), presented in Tucson, AZ, pages 75--85, ACM, ISSN: , ISBN: 978-1-4503-0102-2, Jun. 2011, )
AbstractThe steadily increasing number of nodes in high-performance computing systems and the technology- and power-constraints in networking lead to sparse large-scale networks. Efficient mapping of application communication patterns to such sparse topologies gains importance as systems grow to petascale and beyond. Such topology mappings are supported in parallel programming frameworks such as MPI, but are often not well implemented. We show that the topology mapping problem is NP-complete and analyze and compare different practical topology mapping heuristics. We demonstrate an efficient and fast new heuristic which is based on graph similarity and show its utility with application communication patterns on real topologies. Our mapping strategies support heterogeneous networks and show significant reduction of congestion on torus, fat-tree, and the PERCS network topologies for irregular problems. We also demonstrate that the benefits of topology mapping grow with the network size and show how our algorithms can be used in a practical setting to optimize communication performance. We argue that maximum congestion and average dilation are good metrics for application performance and network power consumption, respectively. Our efficient topology mapping strategies are shown to reduce network congestion by up to 80%, reduce average dilation by up to 50%, and improve benchmarked communication performance by 18%.