ASPECTS OF COMPUTATION
ON ASYNCHRONOUS
PARALLEL PROCESSORS

Proceedings of the IFIP WG 2.5 Working Conference on
Aspects of Computation on Asynchronous Parallel Processors
Stanford, CA, USA, 22-26 August, 1988

edited by

Margaret WRIGHT
AT&T Bell Laboratories
Murray Hill
New Jersey, U.S.A.

NORTH-HOLLAND
AMSTERDAM - NEW YORK - OXFORD - TOKYO

			CONTENTS

EFFECTIVE SCIENTIFIC APPLICATIONS, Session 1
Chair: Lloyd D. Fosdick
Discussants: Charles L. Lawson, John R. Rice

Solving Elliptic Equations on the Cedar Multiprocessor   	  1
E. Gallapoulos and A. Sameh
 
Parallel Eigenvalue and Singular Value Algorithms for 
Signal Processing						 13
K. V. Fernando and S. J. Hammarling


EFFECTIVE SCIENTIFIC APPLICATIONS, Session 2
Chair: John K. Reid
Discussants: Thomas J. Aird and Richard J. Hanson
 
Analysis of a Two-Level Asynchronous Algorithm for PDES		 23
John R. Rice and Dan C. Marinescu

Parallelization on an MIMD Computer with Real-Time Scheduler,
Gauss-Jordan Example    					 35
Serge G. Petiton

A Vector and Parallel Implementation of the FFT Algorithm 
on the IBM 3090 						 45
Ramesh C. Agarwal
 
A Language Comparison for Scientific Computing on 
MIMD Architectures						 55
Mark T. Jones, Merrell L. Patrick and Robert G. Voigt

EFFECTIVE SCIENTIFIC APPLICATIONS, Session 3
Chair: T. J. Dekker
Discussants: Sven J. Hammarling and Maurice Clint

Multilevel Parallel Solver for Banded Linear Systems    	 69
Ibrahim N. Haj and Stig Skelboe

Asynchronous Parallel Iterative Methods				 79
Avi Lin

A Project for Developing a Linear Algebra Library
for High-Performance Computers       				 87
J. Demmel, J. J. Dongarra, J. DuCroz, A. Greenbaum,
S. J. Hammarling and D. C. Sorensen
 
Monte Carlo Simulations of Lattice Gauge Theories		 93
Philippe de Forcrand

 
LANGUAGES AND PROGRAMMING ISSUES
Chair: Mladen A. Vouk
Discussants: Morven Gentleman and Fred Gustavson

Asynchronous Parallel Execution of a Multiply Nested For Loop
by Vector Indirect Addressing   				101
Takao Tsuda

Ultra-High Performance Multiprocessor Supersystems
with Programmable Architecture   				111
A. V. Kalyaev

LIBRARIES, ENVIRONMENTS AND TOOLS
Chair: W. J. Cody, Jr.
Discussants: Paul C. Messina, Olin G. Johnson

Tools and Methodology for Programming Parallel Processors    	125
Jack J. Dongarra, Danny C. Sorensen and Orlic Brewer

Monitoring Parallel Programs on Message-Passing Multiprocessors 139
A. P. W. Boehm, J. R. Gurd and M. C. Kallstrom

Numerical Library Facilities for Transputer Arrays		159
L. M. Delves and N. G. Brown

Aspects of the Design of Portable and Efficient Basic 
Linear Algebra in Ada						171
M. J. Eri, G. S. Hodgson and L. Stecriman-Clark

LANGUAGE DESIGN ISSUES
Chair: Bo Einarsson
Discussants: Stuart Feldman and Danny C. Sorensen

Automatic MIMD Parallelization 					181
Hans P. Zima

Parallel Programming Experiments with Linda       		193
Nicholas Carriero, David Gelernter, Susanne Hupfer, 
James Narem and Paman Sundaresh
 
Parallel Processing based on Active-Data for Asynchronous 
Systems								201
Chris Jesshope
 
SISAL Implementation and Performance     			211
Stephen Skedzielewski

OPEN SESSION
Chair: John K. Reid
Discussants: Thomas J. Aird, Elias Houstis, Theodore Papatheodorou
 
A Parallel Implementation of Matrix Multiplication and
LU Factorization on the IBM 3090   				217
Ramesh C. Agarwal and Fred G. Gustavson

A Parallel QR Decomposition Algorithm     			222
Lars Elde'n

Optimizing a Reconfigurable Transputer Array for
Line-of-Sight Communications    				223
J. S. Ward, J. B. G. Roberts and P. Simpson
  
Solution of Ordinary Differential Equations by
Waveform Relaxation Methods      				225
Stig Skelboe
 
Recursive Task Generation   					226
Geoffrey Hunter and Tyler A. Ivanco

DINO-A New Language for Numerical Computation
on Distributed Memory Multiprocessors   			232
Matthew Rosing and Robert B. Schnabel

Mapping SISAL Programs for Efficient Execution
on the Connection Machine    					234
Jack B. Dennis

Parallel Computing Forum (PCF) Fortran				235
Brian T. Smith

A Model of Fault-Tolerant Dataflow Computers			236
Milan Ojstersek and Viljem Zumer

SYSTEM DESIGN ISSUES
Chair: Brian Ford
Discussants: Thomas J. Aird and Brian T. Smith

Synchronous versus Asynchronous Communication
in High Performance Multicomputer Systems 			239
P. M. Behr, W. K. Giloi and W. Schroeder

EFFECTIVE SCIENTIFIC APPLICATIONS, Session 4
Chair: George Paul, Jr.
Discussants: Lloyd D. Fosdick and Frederic N. Ris
 
Distributable Algorithms for the Symmetric Eigenproblem		249
M. Clint, D. Roantree and A. Stewart

Coarse Grid Acceleration of Some Domain Decomposition Methods
on Multiprocessors  						255
Garry Rodrigue and Ted Ferretta

Future Scientific Programming					261
Arvind

PARTICIPANTS							265
AUTHOR INDEX 							271