WOPPLOT 83 Parallel processing: Logic, Organization, and Technology: Proceedings of a Workshop Held at the Federal Armed Forces University Munich (HSBw M) Neubiberg, Bavaria, Germany, June 27–29, 1983

199 Pages·1984·5.264 MB·English

by  F. Vester (auth.), Jörg D. Becker, Ignaz Eisele (eds.)

Lecture Notes in Physics
For information about Vols. 1-131, please contact your bookseller or Springer-Verlag.
Vol. 132: Systems Far from Equilibrium. Proceedings, 1980. Vol. 152: Physics of Narrow Gap Semiconductors. Proceed
Edited by L. Garrido. XV, 403 pages. 1980. ings, 1981. Edited by E. Gornik, H. Heinrich and L. Palmets·
hofer. XIII, 485 pages. 1982.
Vol. 133: Narrow Gap Semiconductors Physics and Applica
tions. Proceedings, 1979. Edited by W. Zawadzki. X, 572 Vol. 153: Mathematical Problems in Theoretical Physics.
pages. 1980. Proceedings, 1981. Ediied by R. Schrader, R. Seiler, and
D.A. Uhlenbrock. XII, 429 pages. 1982.
Vol. 134: yy Collisions. Proceedings, 1980. Edited by G.
Cochard and P. Kessler. XIII, 400 pages. 1980. Vol. 154: Macroscopic Properlies of Disordered Media.
Proceedings, 1981. Edited by R. Burridge, S. Childress, and
Vol. 135: Group Theoretical Methods in Physics. Proceed
G. Papanicolaou. VII, 307 pages. 1982.
ings, 1980. Edited by K. 8. Wolf. XXVI, 629 pages. 1980.
Vol. 155: Quantum Optics. Proceedings, 1981. Edited by
Vol. 136: The Role of Coherent Structures in Modelling
C.A. Engelbrecht VIII, 329 pages. 1982.
Turbulence and Mixing. Proceedings 1980. Edited by
J. Jimenez. XIII, 393 pages. 1981. Vol. 156: Resonances in Heavy Ion Reactions. Proceedings,
1981. Edited by K.A. Eberhard. XII, 448 pages. 1982.
Vol. 137: From Collective States to Quarks in Nuclei. Edited
by H. Arenhövel and A. M. Saruis. VII, 414 pages. 1981. Vol. 157: P. Niyogi, Integral Equation Method in Transonic
Flow. XI, 189 pages. 1982.
Vol. 138: The Many-Body Problem. Proceedings 1980.
Edited by R. Guardiola and J. Ros. V, 374 pages. 1981. Vol. 158: Dynamics of Nuclear Fission and Related Collec
tive Phenomena. Proceedings, 1981. Edited by P. David,
T. Mayer-Kuckuk, and A. van derWoude. X, 462 pages.1982.
Vol. 139: H. D. Doebner, Differential Geometrie Methods in
Mathematical Physics. Proceedings 1981. VII, 329 pages. Vol. 159: E. Seiler, Gauge Theories as a Problem ofConstruc·
1981. tive Quantum Field Theory and Statistical Mechanics. V,
192 pages. 1982.
Vol. 140: P. Kramer, M. Saraceno, Geometry of the Time
Oependent Variational Principle in Quantum Mechanics. Vol. 160: Unilied Theories of Elementary Particles. Critical
IV, 98 pages. 1981. Assessment and Prospects. Proceedings, 1981. Edited by
P. Breitenlohnerand H.P. Dürr. VI, 217 pages. 1982.
Vol. 141: Seventh International Conference on Numerical
Vol. 161: lnteracting Bosons in Nuclei. Proceedings, 1981.
Methods in Fluid Dynamics. Proceedings. Edited by W. C. Edited by J.S. Dehesa, J.M.G. Gomez, and J. Ros. V, 209
Reynolds and R. W. MacCormack. VIII, 485 pages. 1981. pages. 1982.
Vol. 142: Recent Progress in Many-Body Theories. Pro Vol. 162: Relativistic Action at a Distance: Classical and
ceedings. Edited by J. G. Zabolitzky, M. de Llano, M. Fortes Quantum Aspects. Proceedings, 1981. Edited by J. Llosa.
and J. W. Clark. VIII, 479 pages. 1981. X, 263 pages. 1982.
Vol. 143: Present Status and Aims of Quantum Electro Vol. 163: J. S. Darrozes, C. Francois, Mecanique des Fluides
dynamics. Proceedings, 1980. Edited by G. Gräff, E. Klempt lncompressibles. XIX, 459 pages. 1982.
and G. Werth. VI, 302 pages. 1981. Vol. 164: Stability of Thermodynamic Systems. Proceedings,
1981. Edited by J. Casas-Vazquez and G. Lebon. VII, 321
Vol. 144: Topics in Nuclear Physics I. A Comprehensive
pages. 1982.
Review of Recent Developments. Edited by T.T.S. Kuo and
S.S.M. Wong. XX, 567 pages. 1981. Vol. 165: N. Mukunda, H. van Dam, L.C. Biedenharn, Rel
ativistic Models of Extended Hadrons Obeying a Mass-Spin
Vol. 145: Topics in Nuclear Physics II. A Comprehensive Trajectory Constraint. Edited by A. Böhm and J.D. Dollard.
Review of Recent Developments. Proceedings 1980/81. VI, 163 pages. 1982.
Edited by T. T. S. Kuo and S. S. M. Wong. VIII, 571-1.082
Vol. 166: Computer Simulation of Solids. Edited by C.R.A.
pages. 1981.
Catlow and W.C. Mackrodt. XII, 320 pages. 1982.
Vol. 146: B. J. West, On the Simpler Aspects of Nonlinear Vol. 167: G. Fieck, Symmetry of Polycentric Systems. VI,
Fluctuating. Deep Gravity Waves. VI, 341 pages. 1981.
137 pages, 1982.
Vol. 14 7: J. Messer, Tamperature Dependent Thomas-Fermi Vol.168: Heavy-lon Collisions. Proceedings, 1982. Edited by
Theory. IX, 131 pages. 1981. G. Madurga and M. Lozano. VI, 429 pages. 1982.
Vol. 148: Advances in Fluid Mechanics. Proceedings, 1980. Vol. 169: K. Sundermeyer, Constrained Dynamics. IV, 318
Edited by E. Krause. VII, 361 pages. 1981. pages. 1982.
Vol. 170: Eighth International Conference on Numerical
Vol. 149: Disordered Systems and Localization. Proceed
Methods in Fluid Dynamics. Proceedings, 1982. Edited by
ings, 1981. Edited by C. Castellani, C. Castro, and L. Peliti.
E. Krause. X, 569 pages. 1982.
XII, 308 pages. 1981.
Vol. 171: Time-Oependent Hartree-Fock and Beyond. Pro
Vol. 150: N. Straumann, Allgemeine Relativitätstheorie und
ceedings, 1982. Edited by K. Goeke and P.-G. Reinhard. VIII,
relativistische Astrophysik. VII, 418 Seiten. 1981.
426 pages. 1982.
Vol. 151: lntegrable Quantum Field Theory. Proceedings, Vol. 172: lonic Liquids, Mollen Salts and Polyelectrolytes.
1981. Edited by J. Hietarinta and C. Montonen. V, 251 pages. Proceedings, 1982. Edited by K.-H. Bennemann, F. Brouers,
1982. and D. Quitmann. VII, 253 pages. 1982.
Lecture Notes
in Physics
Edited by H. Araki, Kyoto, J. Ehlers, München, K. Hepp, Zürich
R. Kippenhahn, München, H. A. Weidenmüller, Heidelberg
and J. Zittartz, Köln
196
WOPPLOT 83
Parallel Processing: Logic,
Organization, and Technology
Proceedings of a Workshop Held at the
Federal Armed Forces University Munich (HSBw M)
Neubiberg, Bavaria, Germany, June 27-29, 1983
Edited by J. Becker and I. Eiseie
Springer-Verlag Berlin Heidelberg GmbH 1984
Editors
Jörg O. Becker
Ignaz Eiseie
Institute of Physics, Faculty of Electrical Engineering
Federal Armed Forces University Munich
0-8014 Neubiberg
ISBN 978-3-662-38759-7 ISBN 978-3-662-39650-6 (eBook)
DOI 10.1007/978-3-662-39650-6
This work is subject to copyright. All rights are reserved, whether the whole or part of the material
is concerned, specifically those of translation, reprinting, re-use of illustrations, broadcasting,
reproduction by photocopying machine or similar means, and storage in data banks. Under
§ 54 of the German Copyright Law where copies are made for other than private use, a fee is
payable to ·Verwertungsgesellschaft Wort", Munieh.
© by Springer-Verlag Berlin Heidelberg 1984
Originally published by Springer-Verlag Berlin Heidelberg New York in 1984.
2153/3140-543210
Preface
VDPPIDI' 83 - a WOrkshop on Parallel Processing: Logic, Organization
and Technology - was held on the carrpus of the Federal Anred Forces
University .M.mich.
WOpplot was meant to yield an opportunity to look and talk across
the fences of the various and all too separate fields of research,
sounding p::>ssible lines of developrent in parallel processing.
The workshop was concentrating 6n
- fhysical and technological foundations,
- architectures and algorithrns, and
- cybernetic aspects of organization.
The interdisciplinaiy character of the workshop also led to the
inclusion of several talks which do not have an irrmediate irrpact
on parallel processing but which might have sorre significance in the
future.
The "Logic" called for in our invit ation does not shav up appro
priately in these proceedings. We decided, havever to leave the name
of the neeting as it is, wishing IDPPIDI' to rernain an activ e nucle
ation site.
IV
It is a pleasure to thank all twenty-three participants fran five
European countries and fran Japan for their cx:nting and for the live
ly discussions. In particular the editors would like to thank all
the speakers for their carefully prepared talks and for their manus
cripts.
For all those who rnade the technicalities of the rreeting run so
srroothly we wish to ackn<:Mledge Mrs. E. Göppel, Mr. B. von Harnstein,
and Mr. H. Pechtold.
Finally, particular gratit ude for cooperation and financial support
is due to our sponsors:
Sienens AG, Munich
Masserschmit t-Boel.kCM-Blohm GmbH, Ottobrunn
Freundeskreis der Hochschule der Bundeswehr München, Neuhilierg
Neubiberg, Decerrber 1983
J. Becker I. Eisele
C 0 N T E N T S
P~face
Cybemetic Principles of Organization
F. Vester ••••••••.••••...•••••.•..•.•....•....••••••.•••••••••••••.•.••
Physical and Technological Restrietions of VISI
I. Eisele • • • • • • • • . . . . • • • • • • . • • • • • • • • • . . . • . • • • . • • . • • • • . . . . • • • • . • • . • . . • • • 23
Design Strategies for VISI Logic
E. Hörbst, K. Hominger, G. Sandweg • . • • • • • • • • • . . • . • • • • • . • • . • . • • • . • • • • • • 39
Charge Storage and Charge Transfer in Dynamic .Merrories
J .D. Becker • • . • • • • . • • . • • . • • • • • • • • • • • • • • • • • • • . . • • . • • . . • . • • • . • • • • • • . • • • • • 53
Interpretation and Tirre
G. Quadranti .•.•••.•.•••••.••..•..••••..•.••••.•••.•••••••.••••, . • • • • • • . 69
A Stochastic M::ldel of 1/f Noise and Its Application to Se:niconductors
F • Grüneis • . . . . . . • • • • • • • . . . • • • . • • • . • • • . • • • • • • • . . • • . • • . • . • • • • • • • • • • . • • • . 80
Non Detenninistic Machines and Their Generalizations
A. Bertoni , G. Mauri , N. Sabadini • . . • • • • • • • . • • • • • • • • . . • • • • . • • • • • . . • . • • • 86
Parallel Algorithrrs - Theory and Limitations
P. Weidner and F. Roßfeld • • . . • • • • • • • . • • . . • • • . • • • • • • • . . • . • • . • • • • • • • • . • • • 98
Interconnection NetiDrks for MIMD Machines
L • Ciminiera, D • Demartini , A. Serra • • • • • • • • . . • • . • • • • • • • • • • • • • • • . • • • . • • 11 0
The Organization of Parallel Processing Machines
A.M. Wood •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 132
Organization of Multi-Processor Systerrs for Image Processing
V. Cantoni •.••••••••.••.•••••.••••••••••••••.•••••.•••••••••.•••.•••••• 145
Marnory-coupled Processor Arreys for a Broad Spectr\ml of Applications
G. Fritsch •••••.••.••.•.••....••••••••••••••••....••.••••..•••...•••••• 158
An Experirrental M::ldular Multiprocessor System and Its Kernel Processing Unit
H. Iizuka .••..•••••••••••••••.•..••.••••••••.••.•••••.•••••.••••••.•••• 178
CYBERNETIC PRINCIPLES OF ORGANIZATION
F. Vester
Federal Armed Forces University
Institute for Independence of Technical and Social Change
D-8014 Neubiberg, Federal Republic of Germany
I. INTRODUCTION
In our common language we use quite a lot of associations with other
informatory perceptions than speech: we describe pictures, movements,
give examples, produce feelings. In other words, we try to connect
the speech center in our brain with other cerebral areas like the
visual cortex, the cerebellum with its haptic and motive functions
or the hypothalamus and its emotional interpretations. In our
scientific language we rather try to stick to abstract words, anxious
not to touch those other informational channels.
These channels, however, are as accurately working, as uncorruptible
- or corruptible - grey cells as those which work in the causal-logical
regions of our brain. They have even distinct advantages: visual im
pressions, for instance, are much better in parallel processing. They
enter by simultaneously registering millians of bytes recognizing
pattern in a fraction of a second: a far greater performance than the
linear, sequential processing of words.
As in this paper I shall stress the necess~ty to understand our
environment not as a heterogeneaus list of separate items but as a
pattern, a network of relations, I think it is valuable to use as
manny slides as possible - view taken from reality - to accompany my
talk. This will not only help remembering by a better associative
storage but also help to build up a parallel understanding of what
I say, with additional parts of our brain.
Being a biologist originally, I had to deal with organisms, that is
with open complex systems, from the beginning. And the larger these
systems became - cells, organs, individuals, ecosystems - the more the
common laws of their organization became clear to me. The ecosystem
approach, therefore, seems applicable to all levels of life. Two of
them I will touch in my talk: the level of the brain, as an organ
that reflects already in its hardware the environment it was born in
(by anatomic storage of the first perceptions - until about three
months after birth) and how we can train it (and help it by computer
ized simulation) in the skill of 'pattern recognition'. Secondly, the
2
level of ecosystems, understood as 'organisms' obeing some fundamental
systemic laws of survival. Both acting on each other: the brain by
using its structure to understand and handle the engrammed environment,
and the environment by imprinting directly or indirectly (via science)
our brain. This latter in very different degrees of understanding the
pattern of that environment.
In so far as we act according to this understanding, our environment
is treated differently and so are the ways of our economic strategy
and their effects. Let me therefore first review the effects of our
classical "linear thinking" and of the corresponding planning strategy.
II. THE ISOLATED SYS~ APPROACH AND THE CRISIS OF CIVILIZATION
INTERFERENCE WITll ENVIRONMENT
Throughout the course of human history,in our believe that the buffer
capacity of nature is unlimited, its realm of resources is endless and
that what is technically possible is also achievable - with these
illusions we have rather carelessly interfered with a - hithereto
perfectly - functioning and evolving system, our biosphere, in that
we have gradually imposed more and more artificial systems upon it:
Factories, power stations, and large-scale farmings.
Housing estates, reservoirs, road networks, bridges,and ports.
Natural landscapes, such as those in North America, often had
to give way to sprawling urban systems within only a few decades.
We grafted all these systems- be it traffic networks or theAswan High
Dam with all its implications - on our biosphere in the assumption
that their interaction, their communication, would regulate itself
and remain benificial to men who invented them.
REPAIR SERVICE BEHAVIOUR
If anything went wrong, we thought, all that would be needed to repair
matters would be the application of sufficient technology and of
sufficient energy to the problem. Thus we thought we would be able to
make good any deficiencies in the purity of the air, the fertility of
the soil, energy supplies, the natural water balance or in the health
sector. Confident, that there nothing we could not repair, we neither
cared whether these artificial systems would be viable if left to
themselves nor whether thei.r interactions could manage the load, nor
whether they could be linked with others to form a properly functioning
3
unit.
And still, day by day, we initiate further development projects and
set them between existing systems, without even knowing that we are
dealing with systems, without knowing that anything such as laws of
systems behaviour exist, and rules that determine the chances of
survival of a system.
On this basis we also exert a strange behaviour towards problems. In
our times of critical constellations and crisis one can observe a set
back of constructive tasks while the attention of our decision-makers
is directed more and more towards obvious defects and nuisances and
their activity towards their repair, or elimination. A defect in an
open system however is something very different from a defect in a
machine. There one can exchange parts or glue something tagether and
the machine is working again. In an open system any repair which is
concentrated on the defect part only, will have side-effects, thus
forcing us very soon to repair the repair and to repair the repair of
the repair and so on.
The result is, that we slided inexorably into our actual repair
service behaviour which creates only new problems, is more and more
expensive, makes reasonable planning impossible and leaves us lagging
constantly behind general development. What we need is not to repair
the defects nor to forecast it and when they will occur (another
illusion!) but to create constellations which give those defects
less chance to destabilize our system in consideration.
As lang as we do not understand this, we are faced with the fact that
many of our interferences with the environment which were designed for
quick profits will first affect the environment, then the quality of
our own lives and, as a final effect, will become great economical
problems too - especially in developing countries. There are enough
examples of this.
SIX ASPECTS OF OUR INABILITY TO DEAL WITH COMPLEX SYSTEMS
Firstly: energy-squandering consumer goods and the growth of private
transportation make us and our sub-systems more and more dependent on
energy. The same occurred through the increasing use of energy-intensive
materials and production processes. Such a course of development in
energy use is no progress, because it threatens the stability of the
system.