Analysis of a 3D toroidal network for a shared memory architecture

Authors:
F. Pittelli;D. Smitley
Affiliations:
Supercomputing Research Center, 4380 Forbes Blvd., Lanham, MD;Supercomputing Research Center, 4380 Forbes Blvd., Lanham, MD
Venue:
Proceedings of the 1988 ACM/IEEE conference on Supercomputing
Year:
1988

Citing 1
Cited 10

Deadlock-Free Message Routing in Multiprocessor Interconnection Networks

IEEE Transactions on Computers

The horizon supercomputing system: architecture and software

Proceedings of the 1988 ACM/IEEE conference on Supercomputing
A processor architecture for horizon

Proceedings of the 1988 ACM/IEEE conference on Supercomputing
Performance prediction for the horizon super computer

Proceedings of the 1988 ACM/IEEE conference on Supercomputing
HORSE: a simulation of the horizon supercomputer

Proceedings of the 1988 ACM/IEEE conference on Supercomputing
The fast fourier transform and sparse matrix computations: a study of two applications on teh HORIZON supercomputer

Proceedings of the 1988 ACM/IEEE conference on Supercomputing
An overview of supertoroidal networks

SPAA '91 Proceedings of the third annual ACM symposium on Parallel algorithms and architectures
An effective synchronization network for hot-spot accesses

ACM Transactions on Computer Systems (TOCS)
The Tera computer system

ICS '90 Proceedings of the 4th international conference on Supercomputing
Lower Bounds on Communication Loads and Optimal Placements in Torus Networks

IEEE Transactions on Computers
Packet Synchronization for Synchronous Optical Deflection-Routed Interconnection Networks

IEEE Transactions on Parallel and Distributed Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper describes a synchronized network model suitable for the Horizon architecture. The model is defined in terms of a topology and routing policy. A three dimensional toroidal topology is investigated for its multiple redundant paths, memory locality and simplicity in both routing and construction. The routing policy is based on a desperation routing scheme in which it is not guaranteed that a message will make progress on a given network cycle. This scheme requires no complex deadlock avoidance algorithms or node to node flow control, allowing a very simple and efficient implementation. Design considerations and preliminary performance results are discussed.