By Topic

Parallel and Distributed Processing, 1990. Proceedings of the Second IEEE Symposium on

Date 9-13 Dec. 1990

Filter Results

Displaying Results 1 - 25 of 140
  • Proceedings of the Second IEEE Symposium on Parallel and Distributed Processing 1990 (Cat. No.TH0328-5)

    Publication Year: 1990
    Request permission for commercial reuse | PDF file iconPDF (52 KB)
    Freely Available from IEEE
  • Forward chaining parallel inference

    Publication Year: 1990, Page(s):455 - 462
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (700 KB)

    The paper describes the completed and ongoing efforts of the Parallel Inferencing Performance Evaluation and Refinement project (PIPER). The PIPER inference engine is based on the Merit Enhanced Traversal Engine (METE) algorithm which is an extension of Forgy's (1979) RETE algorithm. To evaluate the efficacy of this design and implementation, an iterating 108 rule knowledge base was composed. This... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Accelerated learning on the connection machine

    Publication Year: 1990, Page(s):448 - 454
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (504 KB)

    The complexity of most machine learning techniques can be improved by transforming iterative components into their parallel equivalent. The parallel architecture of the Connection Machine provides a platform for the implementation and evaluation of parallel learning techniques. The architecture of the Connection Machine is described along with limitations of the language interface that constrain t... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Modeling distributed termination with pre-defined partial termination ordering

    Publication Year: 1990, Page(s):373 - 376
    Cited by:  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (300 KB)

    Distributed termination convention (DTC) adopted by most of the concurrent programming languages, is attractive because the programmer is relieved of the control details to achieve termination of distributed programs. This model, however, is not very convincing from the implementation point of view as the communication overhead involved in achieving termination is enormous, especially in a distrib... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parallel time complexity of a heuristic algorithm for the k-center problem with usage weights

    Publication Year: 1990, Page(s):254 - 257
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (292 KB)

    The authors analyze the parallel time-complexity of a heuristic algorithm for solving the NP-hard k-center problem with usage weights. The problem is to choose k vertices as service centers so that the maximum weighted service delivery distance to any vertex is minimized. The algorithm uses a greedy strategy to choose the vertex with maximum usage weights as the next service vert... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On linear speedup of a class of neighborhood functions in an array processor

    Publication Year: 1990, Page(s):441 - 446
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (444 KB)

    Linear speedup of two-dimensional neighborhood functions requires that the underlying processor perform iterative operations and solve the window border problem simultaneously. The authors illustrate a solution to the window border problem using an array processor that provides conflict-free access and alignment of two types of square block vectors of two-dimensional arrays. They give a parallel a... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multiple instruction streams in a highly pipelined processor

    Publication Year: 1990, Page(s):182 - 189
    Cited by:  Patents (15)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (672 KB)

    In a highly pipelined processor, instruction dependencies involving both and control information often limit its potential performance. A cyclic pipeline machine allows multiple instruction streams to share these pipeline stages in time to remove the data and control dependencies. These multiple instruction streams exploit more parallelism in parallel programs. It provides an alternative architect... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Datarol: a massively parallel architecture for functional languages

    Publication Year: 1990, Page(s):726 - 735
    Cited by:  Papers (12)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (732 KB)

    Proposes a parallel machine architecture which incorporates an ultra-multiprocessing facility for parallel execution of functional programs. The machine performs parallel executions along a multi-thread control flow called datarol. A datarol program, instead of using a program counter, the instructions to be executed next are explicitly specified in the preceding instructions. The explicitly speci... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Design of the PRAM network

    Publication Year: 1990, Page(s):367 - 372
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (528 KB)

    The Princeton University PRAM computer network provides process to process data transmission rates equivalent to memory access rates with negligible latency penalties and currently costs less than $1000.00 per gigabit/sec. of network bandwidth. The PRAM network architecture described scales to very large computer networks using fast optical technology. The cost of such a PRAM network implementatio... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Integer sorting on a mesh-connected array of processors

    Publication Year: 1990, Page(s):250 - 253
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (352 KB)

    Schnorr and Shamir and independently Kunde, have shown that sorting N=n2 inputs into snake-like ordering on a n×n mesh requires 3n-o(n) steps. Using a less restrictive, more realistic model the author shows that the sorting N=n 2 integers in the range [1. . .N] can be performed in ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Near-optimal heuristics for schedulings on task-dependent machines

    Publication Year: 1990, Page(s):842 - 847
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (476 KB)

    The paper considers task assignments in a parallel processing environment. Processors in this environment are considered to be task-dependent in that a task's processing time depends not only on the nature of the task but also on the processor to which it is assigned. The problem is a generalization of several related problems of scheduling independent tasks. The authors show that the problem is N... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Convergence/response tradeoffs in concurrent systems

    Publication Year: 1990, Page(s):188 - 192
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (356 KB)

    A self-stabilizing system is one which if started at any unsafe state, is guaranteed to converge to a safe state within a finite number of state transitions. The convergence span of such a system is defined as the maximum number of critical transitions that can be executed before the system reaches a safe state. The authors discuss the tradeoff between the convergence span of a self-stabilizing sy... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A universal online bit-serial cell for parallel expression evaluation

    Publication Year: 1990, Page(s):335 - 338
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (268 KB)

    The paper gives a short description of an arithmetic cell that can perform a variety of online operations on bit-streams representing different codings of numbers. The results of simulations of these cells in the evaluation of composite arithmetic expressions are discussed View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multi-version memory: software cache management for concurrent B-trees

    Publication Year: 1990, Page(s):650 - 655
    Cited by:  Papers (9)  |  Patents (3)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (540 KB)

    The authors describe a new concurrent B-tree algorithm. The algorithm is designed to work well in large-scale parallel or distributed systems in which the number of processors sharing the tree is large or the communication delay between processors (or between processors and the global memory for a shared-memory system) is large relative to the speed of local computation. The basis of the algorithm... View full abstract»

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

    Publication Year: 1990, Page(s):497 - 500
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (196 KB)

    Recurrence formulations for various problems such as finding an optimal order of matrix multiplication, finding an optimal binary search tree and optimal triangulation of polygons assume a similar form. A CREW PRAM algorithm has been given to solve such dynamic programming problems. The algorithm uses O(n6/log n ) processors and runs in O(log... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Recognition and restoration of periodic patterns with recurrent neural network

    Publication Year: 1990, Page(s):436 - 440
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (340 KB)

    Using the fully recurrent network with the temporal supervised learning algorithm developed by Williams and Zipser, the author performed several experiments aimed at recognizing and restoring periodic patterns. The results can be summarized as follows: the recurrent network could recognize complex and multiple patterns simultaneously, if appropriate number of hidden units are given; the network co... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Modeling the execution of LOTOS specifications by cooperating extended finite state machines

    Publication Year: 1990, Page(s):780 - 785
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (592 KB)

    The paper proposes an abstract execution model for LOTOS specifications, which is based on a number of cooperating extended finite state machines, applies to whatever LOTOS specification, with the only exclusion of those containing non well guarded recursion, and is suitable for a distributed implementation. Every machine represents a process, which does not necessarily correspond to one of the pr... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Type inference and the Algebra of qualified relations

    Publication Year: 1990, Page(s):352 - 356
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (360 KB)

    The algebra of qualified relations proposed by Ceri and Pelagatti supports pruning of distributed database queries, minimizing the number of sites involved in evaluation. Domain interval typing in the framework of formal type inference subsumes this approach, pruning not only unsatisfiable subqueries involving fragmentation predicates, but also unsatisfiable subqueries mentioning any relation attr... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A broadcast/reduce architecture for high-speed data compression

    Publication Year: 1990, Page(s):174 - 181
    Cited by:  Papers (7)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (640 KB)

    The author presents a parallel architecture for high-speed data compression based on textual substitution using a sliding window. The architecture combines a systolic array with trees for data broadcast and reduction. Compression involves two steps. First, a match generator computes in parallel the maximal matches available at each position of the input. The generator uses a systolic array to hold... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fault tolerant distributed shared memory algorithms

    Publication Year: 1990, Page(s):719 - 724
    Cited by:  Papers (21)  |  Patents (6)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (604 KB)

    Distributed shared memory (DSM) has received increased attention as a mechanism for interprocess communication in loosely-coupled distributed systems because of its perceived advantages over direct use of message passing or remote procedure calls. One problem with most DSM algorithms proposed to date, however, is that they do not tolerate faults. The paper extends four basic DSM algorithms to tole... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fault tolerant distributed computing using atomic send-receive checkpoints

    Publication Year: 1990, Page(s):215 - 222
    Cited by:  Papers (4)  |  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (548 KB)

    The paper presents a deadlock free fault recovery algorithm for an entirely distributed system in which the messages do not need to arrive in the order they have been sent. The method is based on the asynchronous, atomic checkpointing of the sender and receiver of a message. Messages not balanced in the last permanent checkpoints are recorded in the new checkpoints. The fault recovery is based on:... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multisensor fusion using neural networks

    Publication Year: 1990, Page(s):812 - 815
    Cited by:  Papers (1)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (296 KB)

    A multisensor system for robot navigation has been developed that uses artificial neural networks to perform sensor data fusion. Four neural networks were investigated regarding their potential to fuse data from an ultrasonic and an infrared range finder to yield more accurate estimates of depth. A radial basis predicter using localized receptive fields (LRF) was able to learn mappings quickly. Ho... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A topological property of hypercubes: node disjoint paths

    Publication Year: 1990, Page(s):532 - 539
    Cited by:  Papers (11)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (560 KB)

    A graph G=(V,E) (|V|⩾2n) satisfies property Pn if given any 2n distinct vertices s1,s 2, . . ., sn, g1,g2, . . ., gn V, there exist n node disjoint paths p i[... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Main memory database recovery algorithms and their performance

    Publication Year: 1990, Page(s):152 - 155
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (312 KB)

    In main memory systems recovery becomes more complex mainly due to the volatility of main memory and the elimination of the separation of data storage (disk) and data processing locations (main memory). In a disk based database system, it is the disk copy that represents the stable database restored after a system failure; but in MDBSs, the main memory copy of the database must be recovered. The p... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the fault diagnosis of a VLSI cellular array

    Publication Year: 1990, Page(s):690 - 697
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (480 KB)

    A new cellular array called L-net for implementing fault-tolerant permutation network is presented. Synthesis of this network is accomplished using 6-terminal, 4-state elementary cells. Inter-cell links are limited to between near-neighbor cells only, hence is very suitable for VLSI implementation. The network control is easy. Fault diagnosis and fault tolerance also can be achieved on the L-net. ... View full abstract»

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