Abstract
The increase of interconnecting wires is one of the serious problems of implementing a parallel processing system having over ten thousand processors. We have proposed a cluster architecture called a TRB (Torus-Ring-Bus) for these massive parallel systems. This architecture aims at both to decrease the interconnecting wires and increase the performance of numerical computations. In this paper, we describe the prototype system which consists of 48 processors, and evaluate the network performance, measuring the results with the Red-Black-SOR method.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
References
Intel Corp., Paragon XP/S product overview, 1991.
Cray Research Japan Ltd., CRAY TERA-3D, 1992.
Parsytec, Persytec GC technical summary, 1991.
M. Kohata,T. Itoh,Y. Miyagaki, “ Torus-Ring-Bus Connected Highly Parallel Processor ”, IEEE Pac. Rim. Conf. on Comm., Comp. and Signal Processing, Victoria, 1993.
Texas Instruments, TMS320C30 users manual, 1992.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1996 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Itoh, T., Kohata, M. (1996). Torus-ring-bus connected hierarchic parallel machine. In: Liddell, H., Colbrook, A., Hertzberger, B., Sloot, P. (eds) High-Performance Computing and Networking. HPCN-Europe 1996. Lecture Notes in Computer Science, vol 1067. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-61142-8_622
Download citation
DOI: https://doi.org/10.1007/3-540-61142-8_622
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-61142-4
Online ISBN: 978-3-540-49955-8
eBook Packages: Springer Book Archive