By Topic

Design and Application of Parallel Digital Processors, 1991., Second International Specialist Seminar on the

Date 15-19 Apr 1991

Filter Results

Displaying Results 1 - 25 of 31
  • Text retrieval on the AMT DAP

    Publication Year: 1991, Page(s):105 - 111
    Cited by:  Patents (1)
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (376 KB)

    Describes the use of massively parallel computing (in particular, the AMT DAP series) for word based text document retrieval. Wider utilisation of ever larger databases requires higher throughputs of queries scanning larger volumes of data. Conventional systems cannot cope with thousands of queries/minute covering tens of GBytes of data; a single query term may invoke a million hits, and even for ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The benefits of database computers

    Publication Year: 1991, Page(s):112 - 117
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (388 KB)

    Parallel processing has long been recognized as a highly efficient method of applying large amounts of computing power to specialised applications. The world of commercial computing however, has for the most part, failed to supply suitable work to justify the creation of commercial parallel systems. The advent of the relational model, with its parallel language-SQL-and the need to hold large amoun... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A systolic array for RLS and CRLS estimation

    Publication Year: 1991, Page(s):132 - 136
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (232 KB)

    The linear least squares problem is one of the oldest problems in applied mathematics, with numerous applications throughout modern engineering. In the traditional information matrix approach to recursive least squares (RLS) estimation, a triangular factor R of A is updated to a triangular factor R of R_ of A_, by means of orthogonal transformations, and then a triangular backsolve i... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parallelization of a predator-prey simulation using time warp

    Publication Year: 1991, Page(s):137 - 141
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (320 KB)

    It has long been considered that the predator-prey simulation can only be tackled by time-driven simulation because unpredictable interactions may only be detected by exhaustive testing. The authors describe the parallelization of this application using a `Simulation Testbed' developed within the Parallel Systems Group at Exeter University. Mapping a simulation onto a message-passing parallel arch... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Algorithmic engineering in digital signal processing

    Publication Year: 1991, Page(s):11 - 18
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (336 KB)

    The authors introduce the concept of algorithmic engineering and describe the implementation of simple matrix operators using an array of basic processing cells. Algorithmic engineering is best carried out in terms of the basic signal flow graph associated with each elementary operator. The author then considers recursive least squares estimation and a parallel processing architecture for it. This... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Systolic array building blocks for real-time signal processing

    Publication Year: 1991, Page(s):24 - 29
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (324 KB)

    The authors describe the application of a number of contrasting module designs to adaptive processing systems. In particular, they illustrate that, whereas the transputer can effectively implement complex parallel algorithms designed for audio band problems similar algorithms for RF antenna array processing have demanded the development of a special DSP chip offering a much higher processing throu... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Implementations of multilayer neural networks on parallel architectures

    Publication Year: 1991, Page(s):43 - 47
    Cited by:  Papers (1)
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (364 KB)

    The authors study the implementation of multilayered neural networks on two parallel architectures: the Connection Machine, a SIMD type computer, and the TNode, a Transputer based MIMD computer. They compare various implementations and present experimental results and performance evaluations View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Radar vision

    Publication Year: 1991, Page(s):75 - 78
    Cited by:  Papers (1)
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (336 KB)

    The authors discusses a novel idea, for which he has coined the term radar vision (Haykin 1990). The goal is to make radar an intelligent remote sensing device, such that it is capable of developing cognition of the surrounding environment. To illustrate the main points of this new idea, he uses the case of a coherent marine radar for ocean surveillance. The special targets of interest include gro... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Getting parallelism into perspective

    Publication Year: 1991, Page(s):142 - 144
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (168 KB)

    Parallelism raises many novel issues not present in sequential computing, as well as new twists to older issues. How is the ordinary user to come to terms with exploiting parallelism for applications? It is argued that, by developing and characterising the concept of an abstract programming model based on a virtual architecture, a clearer perspective is achieved. Two examples of virtual architectu... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Why should parallelism help design KBS control applications?

    Publication Year: 1991, Page(s):118 - 121
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (280 KB)

    When one starts to design KBS process control applications, one generally faces, whatever the approach taken, a need for a great amount of computer power both numerical and symbolic. The author suggests some directions for potential parallelism in an alternative architecture to blackboard. However some results such as the ones from Esprit project 413, i.e. the natural parallel reduction model and ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The case for custom computation

    Publication Year: 1991, Page(s):60 - 64
    Cited by:  Patents (1)
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (208 KB)

    The paper describes a general purpose custom computer and the architectural styles supported by it. They use a 2×4 array of CAL1024 parts to form the CHS2×4 board. The paper seeks to make the argument that custom computers represent a fundamental innovation in computer architecture by supporting an alternative paradigm for computation, the configurable hardware paradigm, within a gener... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Distributed discrete-event simulations running on multi-transputer arrays

    Publication Year: 1991, Page(s):122 - 126
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (344 KB)

    The two major techniques which are currently being used for parallel discrete-event simulation are the conservative method developed by Chandy and Misra (1971) and the optimistic technique of Jefferson (1985). In the conservative technique a process will only process its event queue if it is certain that it is safe to do so-i.e. it cannot receive an event-notification message referring to its past... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Case studies of realtime multiprocessors in robotics

    Publication Year: 1991, Page(s):1 - 5
    Cited by:  Papers (3)
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (412 KB)

    The National Research Council has conducted research on sensor-based robotics for many years. Three of the many applications built are described here. The first is a motion system controller, capable of coordinating the actions of multiple robot arms. The second is a tracking system, which uses a pair of laser range finders to enable a robot arm to track, and to grasp, an object in motion. The thi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Real-time SAR change-detection on a parallel computer architecture

    Publication Year: 1991, Page(s):19 - 23
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (368 KB)

    Synthetic aperture radar (SAR) provides a coherent, high-resolution, imaging system. The author outlines the principles underlying the implementation of a real-time SAR processor capable of generating linear focussed imagery by combining motion compensation, together with autofocus and phase correction. Finally, the changes between the SAR image and a previous version are detected to indicate sign... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Building parallel analysis software

    Publication Year: 1991, Page(s):30 - 32
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (204 KB)

    The author describes the basic principles of FEA and place that method in its industrial context. He then describes his preconceived ideas as to how the parallel software should be structured, and what problems he found, and the structure which he eventually used. The new approach can be presented as a natural extension of traditional software design, and one of the nicer discoveries was that trad... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Aspects of vertical mode in multiprocessor systems, unconventional computation on conventional processors

    Publication Year: 1991, Page(s):48 - 54
    Cited by:  Patents (5)
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (264 KB)

    The author discusses what typical vertical operations are and the advantages they have. Using 1 bit registers gives a completely variable wordlength. He discusses: matching and searching operations and their vertical efficiency; bit-algorithms; and Chebyshev approximation of transcendental functions (Ln, T-polynomials, differential equation transformations; linear equations for coefficients; compu... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Commonality in architectures

    Publication Year: 1991, Page(s):70 - 74
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (260 KB)

    Multiprocessor architectures are converging. For the present, we urgently need to adopt common standards for message passing programming. For the future, one can expect scalable virtual shared memory machines to dominate. The author discusses: communication strategies; dedicated components; programming environments; and programming. An example listing of a ranking program is given that would requi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Developing a stereo vision system for control of an AGV

    Publication Year: 1991, Page(s):79 - 83
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (308 KB)

    In the course of recent work on the ESPRIT VOILA Vision Research Pilot Project, the GEC-Marconi Hirst Research Centre and the AI Vision Research Unit at the University of Sheffield have established a platform for the development of a machine stereo vision system for the control of a robot vehicle. This platform at the Hirst Research Centre is based on a laboratory scale autonomous guided vehicle, ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • How long is a MegaFlop?

    Publication Year: 1991, Page(s):145 - 149
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (380 KB)

    Investigates the nature of measurement, philosophically and historically. Some qualities of `good' and `bad' metrics are discussed. These are then used to take a critical look at the use and abuse of measurement in the empirical science of the development and promotion of parallel digital processing. The familiar metrics of speed-up, scaled speed-up, MegaFlops, and memory size are critiqued within... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A systolic Broyden algorithm

    Publication Year: 1991, Page(s):127 - 131
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (248 KB)

    A systolic array for solving nonlinear systems of equations using the Quasi-Newton Broyden algorithm is proposed. The design is based on the idea of reducing a single iteration of the method to a number of Schur complements which can be pipelined on a number of Faddeev arrays. The algorithm requires O(n2) cells for a system of n nonlinear equations in n unknow... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Second International Specialist Seminar on the Design and Application of Parallel Digital Processors (Conf. Publ. No.334)

    Publication Year: 1991
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (60 KB)

    The following topics were dealt with: embedded processing and signal processing; simulation and modelling; architectures; pattern recognition and sensor processing; parallel processing environments; design environments; database systems; and symbolic processing View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Virtual shared memory for the ARAM model of computation using the MP1 packet routing chip

    Publication Year: 1991, Page(s):55 - 59
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (300 KB)

    The author shows how virtual shared memory computers may be constructed from the MP1 communications chip for the ARAM computational model. He suggests solutions to the key issues of load-balancing and latency hiding. He favours the ARAM model interface as it provides a low bandwidth interface to the local processor, avoiding a direct coupling to the address lines of the processor and the low-laten... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parallel computation: parallel communication

    Publication Year: 1991, Page(s):38 - 42
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (284 KB)

    The author uses the following result from contemporary computer science: the ability of two-phase randomised routing to support scalable throughput and low delay when routing on hypercubes (even when routing among the plog(p) processes distributed among p processors) together with the existence of message-passing hardware: processors, preferably with efficient process s... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Mapping algorithms onto parallel architectures: time schedules

    Publication Year: 1991, Page(s):65 - 69
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (316 KB)

    Time scheduling is one of the important problems faced in mapping algorithms onto parallel architectures, particularly in real-time fault-tolerant computing. The paper shows how linear schedules can be used to exploit parallelism available in nested-loop programs. It relates linear schedules to known loop transformation techniques for modeling the parallel execution and extracting parallelism of n... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Systolic systems

    Publication Year: 1991, Page(s):6 - 10
    IEEE is not the copyright holder of this material | Click to expandAbstract | PDF file iconPDF (328 KB)

    Methods for synthesizing systolic arrays from uniform DAGs are well understood. The authors first briefly review the synthesis method using Gaussian elimination as an example. Then they focus upon three recent extensions: DAG uniformization and spacetime minimal arrays; mapping n-dimensional DAGs (n⩽3) onto linear arrays; and partitioning techniques for the efficient mapping of a computational... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.