Content Addressable Memories

Due to continual progress in the large-scale integration of semiconductor circuits, parallel computing principles can already be met in low-cost sys tems: numerous examples exist in image processing, for which special hard ware is implementable with quite modest resources even by nonprofessional designers. Principles of content addressing, if thoroughly understood, can thereby be applied effectively using standard components. On the other hand, mass storage based on associative principles still exists only in the long term plans of computer technologists. This situation is somewhat confused by the fact that certain expectations are held for the development of new storage media such as optical memories and "spin glasses" (metal alloys with low-density magnetic impurities). Their technologies, however, may not ripen until after "fifth generation" computers have been built. It seems that software methods for content addressing, especially those based on hash coding principles, are still holding their position firmly, and a few innovations have been developed recently. As they need no special hardware, one might expect that they will spread to a wide circle of users. This monograph is based on an extensive literature survey, most of which was published in the First Edition. I have added Chap. ?, which contains a review of more recent work. This updated book now has references to over 1200 original publications. In the editing of the new material, I received valuable help from Anneli HeimbUrger, M. Sc. , and Mrs. Leila Koivisto.

Designers and users of computer systems have long been aware of the fact that inclusion of some kind of content-addressable or "associative" functions in the storage and retrieval mechanisms would allow a more effective and straightforward organization of data than with the usual addressed memories, with the result that the computing power would be significantly increased. However, although the basic principles of content-addressing have been known for over twenty years, the hardware content-addressable memories (CAMs) have found their way only to special roles such as small buffer memories and con trol units. This situation now seems to be changing: Because of the develop ment of new technologies such as very-large-scale integration of semiconduc tor circuits, charge-coupled devices, magnetic-bubble memories, and certain devices based on quantum-mechanical effects, an increasing amount of active searching functions can be transferred to memory units. The prices of the more complex memory components which earlier were too high to allow the application of these principles to mass memories will be reduced to a fraction of the to tal system costs, and this will certainly have a significant impact on the new computer architectures. In order to advance the new memory principles and technologies, more in formation ought to be made accessible to a common user.

This book is a collection of the finalized versions of the papers presented at the third Eurographics Workshop on Graphics Hardware. The diversity of the contributions reflects the widening range of options for graphics hardware that can be exploited due to the constant evolution of VLSI and software technologies. The first part of the book deals with the algorithmic aspects of graphics systems in a hardware-oriented context. Topics are: VLSI design strategies, data distribution for ray-tracing, the advantages of point-driven image generation with respect to VLSI implementation, use of memory and ease of parallelization, ray-tracing, and image reconstruction. The second part is on specific hardware, on content addressable memories and voxel-based systems. The third part addresses parallel systems: massively parallel object-based architectures, two systems in which image generated by individual rendering systems are composited, a transputer-based parallel display processor.

Two significant things have happened since the writing of the first edition in 1983. One of them is recent arousal of strong interest in general aspects of "neural computing", or "neural networks", as the previous neural models are nowadays called. The incentive, of course, has been to develop new com puters. Especially it may have been felt that the so-called fifth-generation computers, based on conventional logic programming, do not yet contain in formation processing principles of the same type as those encountered in the brain. All new ideas for the "neural computers" are, of course, welcome. On the other hand, it is not very easy to see what kind of restrictions there exist to their implementation. In order to approach this problem systematically, cer tain lines of thought, disciplines, and criteria should be followed. It is the pur pose of the added Chapter 9 to reflect upon such problems from a general point of view. Another important thing is a boom of new hardware technologies for dis tributed associative memories, especially high-density semiconductor circuits, and optical materials and components. The era is very close when the parallel processors can be made all-optical. Several working associative memory archi tectures, based solely on optical technologies, have been constructed in recent years. For this reason it was felt necessary to include a separate chapter (Chap. 10) which deals with the optical associative memories. Part of its con tents is taken over from the first edition.

Biology of Memory investigates the biological basis of memory and covers topics ranging from short- and long-term post-perceptual memory to memory storage processes, memory microstructures, chemical transfer, and neuronal plasticity. The activity of neuronal networks in the thalamus of the monkey is also examined, along with integrative functions of the thalamocortical visual system of the cat. Comprised of 20 chapters, this book begins with an overview of the mechanisms that inhibit or interfere with short- and long-term memory, followed by a discussion on different retrieval mechanisms for short- and long-term memory. The reader is then introduced to the role of short- and long-term memory in the formation, retention, and utilization of associations, together with the link between memory and the medial temporal regions of the brain. Subsequent chapters focus on anatomical and chemical changes in the brain during primary learning; cellular models of learning and cellular mechanisms of plasticity in Aplysia; trace phenomena in single neurons of hippocampus and mammiliary bodies; and plasticity in single units in the mammalian brain. The book concludes with a description of Occam (Omnium-Gatherum Core Content Addressable Memory), a computer program for a content addressable memory in the central nervous system. This monograph will be useful to biologists, behavioral psychologists, neuropsychologists, neurophysiologists, biophysicists, and biochemists as well as computer scientists and mathematicians.

About the Scope of This Text This book contains two types of material ~ first, the many divergent and often diffuse meanings given to the concepts of association, associative memory, and associative recaZZ are expounded. A review of this kind was felt necessary because there apparently does not exist any single monograph which could serve as a reference to these topics. But the presentation of the main body of this text is motivated by quite other reasons: in recent years, plenty of interesting mathematical and system-theoretical material has been published which makes it possible to gain a view of associative memory which is different from the conventional abstract and computationally oriented approaches. It seems that the basic operation of associative memory, the storage of information together with the relations or links between the data items, and the selective recall of stored information relative to a piece of key or cue information presented, is not restricted to certain computer-technological implementations but can also be reflected in more general mathematically describable processes in certain physical or other systems, especially in their adaptive state changes. It further seems that some generally known forms of associative memory, namely, certain computer technological artifacts, or abstract systems of concepts or data, are in fact special representations of a class of processes characterized as associative memory.