Browse Prior Art Database

Dynamic Architectures: Problems and Solutions

IP.com Disclosure Number: IPCOM000131329D
Original Publication Date: 1978-Jul-01
Included in the Prior Art Database: 2005-Nov-10
Document File: 21 page(s) / 68K

Publishing Venue

Software Patent Institute

Related People

Steven I. Kartashev: AUTHOR [+4]

Abstract

[Figure containing following caption omitted: A standardizable LSI microprocessor with on-chip configuration control directed by operating-system commands promises a series of computer sets dynamically matched to user-program characteristics.] Modern technology opens new dimensions to the designer of complex computers and parallel systems by providing low-cost LSI modules. Since an LSI module now has high computational throughput, it may be equipped with circuits which selectively activate and deactivate its connections with other modules. As a result, a computer architecture assembled from such modules becomes software reconfigurable -- i.e., it may assume several architectural configurations, each of which is characterized by its own topology of activated interconnections between modules. It would be a mistake to assume that LSI technology ";discovered"; software-reconfigurable architectures. They appeared much earlier. What LSI technology did was to make such architectures more economical and to allow reconfigurable interconnections to be used at a deeper level. Heretofore, reconfiguration has only been performed on the level of functional units in multiprocessing systems or between separate computers in multicomputer systems. With the advent of LSI modules, a broader use of reconfigurable interconnections, not only among functional units but also within each unit (on the level of its intermodule connections), may create additional sources of performance improvement in parallel systems. This paper surveys the existing techniques of reconfiguration in parallel systems. It shows that the introduction of between-module reconfiguration may result in the appearance of new architectures, called dynamic architectures, which form in one system an alterable number of computers of variable size. Finally we introduce the organization of a dynamic architecture which may be implemented using one universal LSI module and standard memory units. Also shown are the basic principles of operating systems for such architectures.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 5% of the total text.

Page 1 of 21

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

This record contains textual material that is copyright ©; 1978 by the Institute of Electrical and Electronics Engineers, Inc. All rights reserved. Contact the IEEE Computer Society http://www.computer.org/ (714-821-8380) for copies of the complete work that was the source of this textual material and for all use beyond that as a record from the SPI Database.

Dynamic Architectures: Problems and Solutions

Steven I. Kartashev

Dynamic Computer Architecture, Inc.

Svetlana P. Kartashev

University of Nebraska-Lincoln

(Image Omitted: A standardizable LSI microprocessor with on-chip configuration control directed by operating-system commands promises a series of computer sets dynamically matched to user-program characteristics.)

Modern technology opens new dimensions to the designer of complex computers and parallel systems by providing low-cost LSI modules. Since an LSI module now has high computational throughput, it may be equipped with circuits which selectively activate and deactivate its connections with other modules. As a result, a computer architecture assembled from such modules becomes software reconfigurable -- i.e., it may assume several architectural configurations, each of which is characterized by its own topology of activated interconnections between modules.

It would be a mistake to assume that LSI technology "discovered" software-reconfigurable architectures. They appeared much earlier. What LSI technology did was to make such architectures more economical and to allow reconfigurable interconnections to be used at a deeper level. Heretofore, reconfiguration has only been performed on the level of functional units in multiprocessing systems or between separate computers in multicomputer systems. With the advent of LSI modules, a broader use of reconfigurable interconnections, not only among functional units but also within each unit (on the level of its intermodule connections), may create additional sources of performance improvement in parallel systems.

This paper surveys the existing techniques of reconfiguration in parallel systems. It shows that the introduction of between-module reconfiguration may result in the appearance of new architectures, called dynamic architectures, which form in one system an alterable number of computers of variable size. Finally we introduce the organization of a dynamic architecture which may be implemented using one universal LSI module and standard memory units. Also shown are the basic principles of operating systems for such architectures.

Our presentation of the concepts of dynamic architecture is on a theoretical level. In an article of reasonable length it is not feasible to describe an actual system in all its details. However, the theory rests upon a foundation of practical work.

Dynamic Computer Architecture, Inc., has completed the design of two projects of a multicomputer system with dynamic architecture. It is...