[1991] Proceedings. The Fifth International Parallel Processing Symposium

30 April-2 May 1991

Filter Results

Displaying Results 1 - 25 of 103
  • Proceedings. The Fifth International Parallel Processing Symposium (Cat. No.91TH0363-2)

    Publication Year: 1991
    Request permission for reuse | PDF file iconPDF (71 KB)
    Freely Available from IEEE
  • Efficient implementations of a class of +or-2/sup b/ parallel computations on a SIMD hypercube

    Publication Year: 1991, Page(s):2 - 9
    Cited by:  Papers (1)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (469 KB)

    The authors identify an important class of parallel computations, called +or-2/sup b/-descend, with an efficient implementation on a hypercube. Given the input A(0:N-1), a computation in this class consists of log N iterations. Iteration b, b=log N-1, . . ., 0, computes the new value of each A(i) as a function of A(i), A(i+2/sup b/) and A(i-2/sup b/). They obtain a general algorithm for implementi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Triangulation in a plane and 3D convex hull on mesh-connected arrays and hypercubes

    Publication Year: 1991, Page(s):10 - 17
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (745 KB)

    The authors present four new parallel algorithms for triangulation of points in a plane. The first two, for one-way iterative arrays and two-way cellular arrays, requiring O(n) time and O(n) processors. Next they give an algorithm for d-dimensional mesh arrays which requires O(n/sup 1/d/) time and a hypercube algorithm with a worst case running time of O(log/sup 3/ n) and an expected running time ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The effects of communication latency upon synchronization and dynamic load balance on a hypercube

    Publication Year: 1991, Page(s):18 - 25
    Cited by:  Papers (2)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (537 KB)

    The effects of communication latency upon dynamic load balance for computations which require global synchronization are discussed. Experimental results from the study of the performance of iteration methods on an NCUBE 1 are presented. Discussions are restricted to the case when all processing elements (PEs) execute the same program, but on different data; the so-called SPMD paradigm.<<ETX&... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Improved graph computations on the reconfigurable mesh

    Publication Year: 1991, Page(s):26 - 29
    Cited by:  Papers (1)  |  Patents (4)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (287 KB)

    The paper develops a data structure and data movement operations which lead to efficient parallel graph computations on the reconfigurable mesh. The technique computes a minimal spanning forest of a N/sup 1/2/ vertex graph in O(log N log log N) time given the adjacency matrix of the graph on a N/sup 1/2/*N/sup 1/2/ reconfigurable mesh. This improves over the known algorithms of O(log/sup 2/ N) tim... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Design guidelines for parallel algorithms using continuous job profiles

    Publication Year: 1991, Page(s):30 - 36
    Cited by:  Patents (7)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (453 KB)

    Due to the increased use of parallelism in computer systems, designers need tools to evaluate job performance and match problems with candidate architectures. In the paper, jobs are modeled using continuous job profiles. Five classes of continuous job profiles are analyzed in terms of the number of processors used under the processor sweeping scheduling discipline. Characteristics of each job clas... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An inherently fault tolerant sorting algorithm

    Publication Year: 1991, Page(s):37 - 42
    Cited by:  Papers (4)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (489 KB)

    The paper defines inherent fault tolerance and illustrates this approach by developing an inherently fault tolerant parallel sorting algorithm. In particular, it shows, how the algorithm can be developed systematically in four steps, namely, by starting with a conventional algorithm, extending it to an infinite iterative algorithm, incorporating inherent fault tolerance, and improving the performa... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Performance analysis of algorithms on a message passing multiprocess

    Publication Year: 1991, Page(s):43 - 50
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (408 KB)

    The paper discusses a model for analysing the performance of algorithms executed on a message passing multiprocessor which achieve reduction in parallel execution time by domain decomposition. The model can estimate the performance of algorithms with different execution behaviours and incorporates the effect of overlapping of computation and communication, algorithmic waiting time and load imbalan... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Routing algorithms in interval and circular-arc networks

    Publication Year: 1991, Page(s):51 - 55
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (330 KB)

    A simple parallel algorithm is shown for routing messages between the nodes of a network whose underlying graph is an interval graph. Each node executing the algorithm makes purely local decisions about where to route the message it receives. The algorithm uses constant message length and shortest-path routing. A straightforward extension of the algorithm allows handling time-varying faults in the... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Deterministic timing schema for parallel programs

    Publication Year: 1991, Page(s):56 - 63
    Cited by:  Papers (8)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (594 KB)

    Formula or schema for computing the best and worst case execution times of a variety of parallel program constructs are presented. Timing schema are proposed for processes that do not interact but synchronize at their start and end, for shared variable interactions through critical sections and general semaphores, and for distributed message passing and remote procedure call. Also included are rep... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Generalized formulation and hypercube algorithms for relaxation labeling

    Publication Year: 1991, Page(s):64 - 69
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (372 KB)

    Presents a generalized formulation for several well-known approaches to relaxation labeling, including discrete, fuzzy, linear probabilistic models and several nonlinear probabilistic modes. Based on this generalized framework, two parallel algorithms for SIMD hypercube computers with different numbers of processors are proposed and analyzed. The algorithms achieve minimal time complexity.<<... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • General asynchrony is not expensive for PRAMs

    Publication Year: 1991, Page(s):70 - 75
    Cited by:  Papers (1)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (515 KB)

    The authors present a technique to efficiently simulate a synchronous n-processor PRAM using a completely asynchronous PRAM. This work is an extension of prior results of Martel et. al. (1990) in which a measure of work for asynchronous PRAM algorithms is developed that accounts for all instructions executed by the processors including busy waiting steps. These prior results show that if the async... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fast parallel algorithms for solving triangular systems of linear equations on the hypercube

    Publication Year: 1991, Page(s):76 - 83
    Cited by:  Patents (1)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (574 KB)

    Presents efficient hypercube algorithms for solving triangular systems of linear equations by using various matrix partitioning and mapping schemes. Recently, several parallel algorithms have been developed for this problem. In these algorithms, the triangular solver is treated as the second stage of Gauss elimination. Thus, the triangular matrix is distributed by columns (or rows) in a wrap fashi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimal EREW parallel algorithms for connectivity, ear decomposition and st-numbering of planar graphs

    Publication Year: 1991, Page(s):84 - 91
    Cited by:  Papers (4)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (633 KB)

    Parallel EREW deterministic algorithms for finding the connected components, ear decomposition and st-numbering of a planar graph are presented. The algorithms take O(log(n)) time with /sub log(n)///sup n+m/ processors. Previous results have the same complexity, but use the CRCW model. The same algorithms can be used for graphs with low genus. Let g be the genus of the minimal embedding of the gra... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Scalability of parallel sorting on mesh multicomputers

    Publication Year: 1991, Page(s):92 - 101
    Cited by:  Papers (9)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (733 KB)

    The paper presents two new parallel algorithms QSP1 and QSP2 based on sequential quicksort for sorting data on a mesh multicomputer, and analyzes their scalability using the isoefficiency metric. It shows that QSP2 matches the lower bound on the isoefficiency function for mesh multicomputers. The isoefficiency of QSP1 is also fairly close to optimal. Lang et al. (1985) and Schnorr et al. (1986) ha... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An algorithm for generating node disjoint routes in Kautz digraphs

    Publication Year: 1991, Page(s):102 - 107
    Cited by:  Papers (9)  |  Patents (2)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (413 KB)

    The authors focus on a particular class of interconnection networks: Kautz networks. These networks have nice properties: a network with degree d and N=d/sup k/+d/sup k-1/ nodes (for any cardinal d, k>0), has a diameter of at most /sup d/log N, the degree d is fixed and independent of the network size. The network is fault-tolerant and the connectivity is d. There is a straightforward mapping f... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A neural-type parallel algorithm for fast matrix inversion

    Publication Year: 1991, Page(s):108 - 113
    Cited by:  Papers (2)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (367 KB)

    The paper introduces the orthogonalized back-propagation algorithm (OBA), a training procedure for adjusting the weights of a neural-type network used for matrix inversion. In this framework the adjustable weights correspond to the estimate of the inverse of the matrix. The algorithm is iterative, in the sense that an initial estimate of the solution is chosen and then updated according to some er... View full abstract»

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

    Publication Year: 1991, Page(s):114 - 119
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (357 KB)

    Two sequences of items sorted in increasing order are given: a sequence A of size n and a sequence B of size m. It is required to determine, for every item of A, the smallest item of B (if one exists) that is larger than it. The paper presents two parallel algorithms for the problem. The first algorithm requires O(logm+logn) time using n processors on an EREW PRAM. On an EREW PRAM with p (p<or=... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • SIMD data communication algorithms for multiply twisted hypercubes

    Publication Year: 1991, Page(s):120 - 125
    Cited by:  Papers (1)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (444 KB)

    The paper explores the effectiveness of multiply-twisted hypercube networks for parallel computing by considering interprocessor communication problems. It presents SIMD parallel data broadcasting, census, and shortest path finding algorithms for multiply-twisted hypercube networks. The data broadcasting algorithms take ((n+1)/2) communication steps to broadcast a message from a processor to all o... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Two EREW algorithms for parentheses matching

    Publication Year: 1991, Page(s):126 - 131
    Cited by:  Papers (2)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (422 KB)

    The authors present two new parallel algorithms for matching parentheses on an exclusive-read exclusive-write parallel random-access machine (EREW PRAM). The first algorithm uses n processors and O(n) space, and requires O(log n) time to match n parentheses. The second algorithm is cost-optimal, and uses O(/sub logn///sup n/) processors and O(n log n) space, and it requires O(log n) time. These al... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A cost-optimal parallel algorithm for the parentheses matching problem on an EREW PRAM

    Publication Year: 1991, Page(s):132 - 137
    Cited by:  Papers (6)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (447 KB)

    The article presents a cost-optimal parallel algorithm for the parentheses matching problem on the EREW PRAM model. For n parentheses, the algorithm requires O(n/p+log n) time and O(n+p log p) space, employing p processors. Thus, for p<or=n/log n, it achieves optimal speedup, requiring O(log n) time and O(n) space. Though the time complexity of the algorithm is comparable with those of the two ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A parallel approximation algorithm for solving one-dimensional bin packing problems

    Publication Year: 1991, Page(s):138 - 143
    Cited by:  Papers (3)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (343 KB)

    Describes a parallel approximation algorithm that can be used to obtain solutions to the one-dimensional bin packing problem: a list L of n items with sizes in the interval View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fully normal algorithms for incomplete hypercubes

    Publication Year: 1991, Page(s):144 - 150
    Cited by:  Papers (2)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (432 KB)

    The paper shows that composite and incomplete hypercubes are algorithmically equivalent. It develops a class of algorithms called fully normal algorithms for incomplete hypercubes with applications in several important fields. Finally it shows that incomplete hypercubes achieve close to linear speedup for FNA over a wide range of the size of the architecture.<<ETX>> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Refining algorithm mappings for linear systolic arrays

    Publication Year: 1991, Page(s):151 - 154
    Cited by:  Papers (1)  |  Patents (1)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (297 KB)

    Nested FOR loop algorithms are the most common applications of systolic arrays. All the techniques for mapping nested FOR loops onto systolic arrays have the 'curse of dimensionality'. The authors propose the approach of first finding coarse-grained mappings that are easier to determine and then refining these mappings through some simple transformations, to obtain efficient fine-grained mappings.... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Analysis of parallel mixed-mode simulation algorithms

    Publication Year: 1991, Page(s):155 - 160
    Cited by:  Papers (2)
    Request permission for reuse | Click to expandAbstract | PDF file iconPDF (467 KB)

    This paper presents a performance analysis of several parallel algorithms for the mixed-mode simulation of digital systems. The performance model is an architectural simulation of the parallel algorithms running on a hypercube multiprocessor. Workload data is collected from a uniprocessor-based mixed-mode simulator on several benchmark circuits, and two distinct circuit partitioning algorithms are... View full abstract»

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