Designing low‐diameter interconnection networks with multi‐ported host‐switch graphs

Abstract

A host-switch graph was originally proposed as a graph that represents a network topology of a computer systems with 1-port host computers and Δ-port switches. It has been studied from both theoretical and practical aspects in terms of the diameter, the average shortest path length, and the performance of real applications. In recent high-performance computing systems, however, a host computer is connected to multiple switches by using InfiniBand, NVSwitch, or Omni-Path, and consequently they provide high bandwidths. Since a host-switch graph cannot represent such systems, this article extends a host-switch graph so that it can represent such systems. As a result, a host-switch graph can include multi-ported hosts. Furthermore, we propose to use multi-port hosts for reducing the diameter. We show that the diameter minimization is equivalent to solving the degree diameter problem for bipartite graphs of diameter three. Our experimental results show that we can drastically reduce the diameter as well as increasing the bandwidth and improves performance of MPI applications by up to 162% as compared with networks with single-ported hosts.

Edge-Based Heuristics for Optimizing Shortcut-Augmented Topologies for HPC Interconnects

“…Another related line of topology research stems from the famous small-world phenomenon, first proposed by [15], that demonstrates that people living in a country constitute societies of a network with only short path lengths. Later, Wattz and Strogatz characterized the small-world phenomenon into the Wattz-Strogatz (WS) model [16], which allows networks to be generated with short average distance and large clustering coefficient [17,18]. The model uses a few additional long edges to reduce the diameter in random graphs for social networks and Internet topologies [3,18].…”

“…Later, Wattz and Strogatz characterized the small-world phenomenon into the Wattz-Strogatz (WS) model [16], which allows networks to be generated with short average distance and large clustering coefficient [17,18]. The model uses a few additional long edges to reduce the diameter in random graphs for social networks and Internet topologies [3,18]. Since then, researchers have been exploring the small-world phenomenon in computer networks [19][20][21].…”

A Novel Interconnection Network with Improved Network Cost Through Shuffle-Exchange Permutation Graph

“…The diameter is the maximum value of the shortest path between any two nodes in the network and is the lower limit of the delay time required to transmit information to the entire network. A network having a relatively small diameter compared to the number of nodes, despite the short distance between nodes, has a disadvantage that it is difficult to design the network in terms of hardware as the number of nodes increases [16][17][18]. The interconnection networks that have been proposed until now can be classified into the following three types according to the number of nodes: mesh class with k × n nodes, hypercube class with 2 n nodes, and star graph class with n!…”