Matrix operations on a multicomputer system with switchable main memory modules and dynamic control

Authors:
S. Y. W. Su;A. K. Thakore
Affiliations:
Univ. of Florida, Gainesville, FL;Univ. of Florida, Gainesville, FL
Venue:
IEEE Transactions on Computers
Year:
1987

Citing 13
Cited 2

Supercomputing

Computer - IEEE Centennial: the state of computing
The cosmic cube

Communications of the ACM - Special section on computer architecture
The Architecture of SM3: A Dynamically Partitionable Multicomputer System

IEEE Transactions on Computers
Ultracomputers

ACM Transactions on Programming Languages and Systems (TOPLAS)
The CRAY-1 computer system

Communications of the ACM - Special issue on computer architecture
Performance evaluation of the statistical aggregation by categorization in the SM3 system

SIGMOD '84 Proceedings of the 1984 ACM SIGMOD international conference on Management of data
B - HIVE: a heterogeneous, interconnected, versatile and expandable multicomputer system

ACM SIGARCH Computer Architecture News
SM3: A Dynamically Partitionable Multicomputer System with Switchable Main Memory Modules

Proceedings of the First International Conference on Data Engineering
Communication Issues in the Design and Analysis of Parallel Algorithms

IEEE Transactions on Software Engineering
Multitasking on the Cray X-MP-2 Multiprocessor

Computer
Multiprocessor Supercomputers for Scientific/Engineering Applications

Computer
A Benchmark Comparison of Three Supercomputers: Fujitsu VP-200, Hitachi S810/120, and Cray X-MP/2

Computer
Supercomputers for Superproblems: An Architectural Introduction

Computer

A linear array of processors with partially shared memory for parallel solution of PDE

Proceedings of the 1990 ACM/IEEE conference on Supercomputing
Experimental analysis of communication/data-conditional aspects of a mixed-mode parallel architecture via synthetic computations

Proceedings of the 1990 ACM/IEEE conference on Supercomputing

Quantified Score

Hi-index	14.98

Visualization

Abstract

This paper presents an analysis and evaluation of the performance of a multicomputer system (SM3) in supporting two basic matrix operations, namely multiplication and inversion. The system supports the efficient execution of the above mentioned operations by 1) achieving a high-bandwidth data transfer among computers by switching main memory modules, 2) supporting network partitioning, 3) employing a hardware communication and synchronization scheme, 4) using a distributed control technique, and 5) providing means to dynamically transfer control. Timing equations are derived and evaluated in an attempt to analyze the performance. Different cases which arise due to the relative sizes of memory modules and matrices during matrix multiplication are analyzed. The cases of partial and maximal pivoting during inversion are also analyzed. The SM3 system is compared quantitatively and qualitatively to a hypercube architecture.