Proceedings of the International Symposium on Parallel Architectures, Algorithms and Networks (ISPAN)

14-16 Dec. 1994

Filter Results

Displaying Results 1 - 25 of 58
  • Proceedings of the International Symposium on Parallel Architectures, Algorithms and Networks (ISPAN)

    Publication Year: 1994
    Request permission for commercial reuse | |PDF file iconPDF (444 KB)
    Freely Available from IEEE
  • A novel hypercube with lower latency

    Publication Year: 1994, Page(s):413 - 420
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (284 KB)

    In this paper, we present a hypercube with reduced diameter and latencies. This is achieved by reconnecting the hypercube with twisted and random connections while keeping the number of wires constant. This new topology is called RT-cube. The RT-cubes offer smaller diameter and reduced latencies compared to the normal hypercubes, resulting in a faster interconnection network topology. A packet swi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Integrated VLSI layout compaction and wire balancing on a shared memory multiprocessor: evaluation of a parallel algorithm

    Publication Year: 1994, Page(s):49 - 56
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (360 KB)

    We first present a unified formulation to three problems in VLSI physical design: layout compaction, wire balancing and integrated layout compaction and wire balancing problem. The aim of layout compaction is to achieve minimum chip width. Whereas wire balancing seeks to achieve minimum total wire length, integrated layout compaction and wire balancing seeks to minimize wire length maintaining the... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Processing nested loop structure with data-flow dependence on a CAM-based processor HAPP

    Publication Year: 1994, Page(s):119 - 126
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (408 KB)

    We know that a significant advantage of content addressable memory (CAM) is that operations are performed locally, thus it can eliminate the problem of bottleneck between processor and memory. In this paper, we propose a CAM-based associative processing processor (HAPP) which is able to combine with a general processor to form an array-processor system, and besides retrieval operations, it can ass... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An interprocessor memory access arbitrating scheme for the S-3800 vector supercomputer

    Publication Year: 1994, Page(s):262 - 269
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (456 KB)

    Reports an instruction-based variable priority scheme which achieves high sustained memory throughput on a tightly coupled multiprocessor (TCMP) vector supercomputer. We analyze the two types of priority control for arbitrating interprocessor memory access conflict. In the case of request level priority control, mutual obstruction causes performance degradation, while in the case of fixed priority... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Increase analysis in the total execution time of a parallel program

    Publication Year: 1994, Page(s):390 - 397
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (356 KB)

    Lower bound on the finishing time of optimal schedules is used as an absolute performance measure of static scheduling heuristics. This paper presents an efficient method of computing such a bound based on estimating overlaps among the execution ranges of tasks in a given task graph and analyzing the delays of tasks on the critical paths of the graph. The computation performed by this method is sh... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parallel connectivity algorithms on permutation graphs

    Publication Year: 1994, Page(s):97 - 104
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (360 KB)

    In this paper, we shall present several algorithms for determining the maximum number of vertex connectivity, testing k-vertex connectivity, determining the maximum number of vertex disjoint s-t paths and finding k-vertex disjoint s-t paths problems on a permutation graph, respectively. We first give several O(n2) time sequential algorithms for determining the maximum number of vertez c... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Distributed validation of massively parallel machines

    Publication Year: 1994, Page(s):326 - 333
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (400 KB)

    In this paper, a distributed algorithm for validating message passing-machines is presented and evaluated. Our approach is based on adaptive distributed diagnosis of multiprocessor systems in a user environment where a full self-diagnosis is not needed. We analyze the algorithm performance using a model based on an open queueing network View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimal parallel algorithm for edge-coloring partial k-trees with bounded degrees

    Publication Year: 1994, Page(s):167 - 174
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (352 KB)

    Many combinatorial problems can be efficiently solved for partial k-trees (graphs of treewidth bounded by k). The edge-coloring problem is one of the well-known combinatorial problems for which no NC algorithms have been obtained for partial k-trees. This paper gives an optimal and first NC parallel algorithm to find an edge-coloring of any given partial k-tree using a minimum number of colors if ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Views of mixed-mode computing and network evaluation

    Publication Year: 1994, Page(s):1 - 8
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (448 KB)

    Trade-offs between the SIMD and MIMD models of architecture for parallelism are presented. Mixed-mode parallelism, where a machine can switch between the SIMD and MIMD modes of parallelism at instruction-level granularity with generally negligible overhead, is discussed. Advantages and disadvantages of mixed-mode parallelism and an example of a mixed-mode parallel algorithm are given. The relation... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Cube-connected modules: a family of cubic networks

    Publication Year: 1994, Page(s):57 - 64
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (340 KB)

    A family of cubic networks, named cube-connected modules, is proposed in this paper. The cube-connected modules network consists of modules which are interconnected as a hypercube. Any connected graph, e.g., cycle, hypercube graph, and complete graph, can serve as a module. Topological properties are investigated, and the problems of routing, broadcasting, embedding, and finding parallel routing p... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Software pipelining for Jetpipeline architecture

    Publication Year: 1994, Page(s):127 - 134
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (392 KB)

    High performance processors based on pipeline processing play an important role in scientific computation. We have proposed a hybrid pipeline architecture named Jetpipeline in our former work. The concept of Jetpipeline comes from the integration of superscalar, VLIW and vector architectures. Jetpipeline has multiple instruction pipelines, which execute multiple instructions like superscalar archi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Efficient implementation techniques for vector memory systems

    Publication Year: 1994, Page(s):270 - 277
    Cited by:  Papers (2)  |  Patents (4)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (360 KB)

    Existing vector machines' memory systems use heavy interleaving and SRAM technology for faster data access. In this paper, we present an efficient vector memory architecture that adopts prime-degree memory interleaving and exploits the capabilities of new-generation DRAM chips with small SRAM cache. The major contribution of this paper is an incremental indexing scheme for prime-degree memory inte... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A greedy task clustering heuristic that is provably good

    Publication Year: 1994, Page(s):398 - 405
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (356 KB)

    A simple greedy algorithm is presented for task clustering with duplication (or recomputation) which, for a task graph with arbitrary granularity, produces a schedule whose makespan is at most twice optimal. Furthermore, the quality of the schedule improves as the granularity of the task graph increases. For example, if the granularity is at least ½, the makespan of the schedule is at most ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An algorithm for maintaining consistent view of processes in distributed systems

    Publication Year: 1994, Page(s):33 - 40
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (392 KB)

    In the paper, the problem of determining the global properties of distributed systems is addressed. At each moment during the execution of a system, every process has its knowledge about the system. By message passing the processes can exchange their knowledge. We present a general algorithm for a process to synthesize the knowledge that it obtains, and to maintain its consistent view about the sy... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parallel graph isomorphism detection with identification matrices

    Publication Year: 1994, Page(s):105 - 112
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (380 KB)

    In this paper, we investigate some properties of identification matrices and exhibit some uses of identification matrices in studying the graph isomorphism problem, a well-known long-standing open problem. We show that, given two m×n identification matrices representing two graphs according to a certain relation, isomorphism can be decided efficiently in parallel if an m×(n-c) submatri... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Message transfer algorithms on the recursive diagonal torus

    Publication Year: 1994, Page(s):310 - 317
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (348 KB)

    Recursive diagonal torus (RDT) is a class of interconnection network for massively parallel computers with 216 nodes. In this paper, message transfer algorithms on the RDT are proposed and discussed. First, a simple one-to-one message routing algorithm called the vector routing is introduced and the result of its practical extension called the floating vector routing is reported. Next, ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Compiler-chosen operator granularity in a functionally-programmed tagged token architecture

    Publication Year: 1994, Page(s):334 - 341
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (384 KB)

    In a recently proposed parallel execution model, operators as well as data operands are dynamically generated at execution time, and enter a special matching store which must accommodate either operator or operands arriving first in time. Here, we extend this execution model to permit the compiler to adjust the granularity of scheduled units emerging from the matching store, a problem which is com... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Transit: reliable high speed interconnection technology

    Publication Year: 1994, Page(s):350 - 357
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (360 KB)

    Important improvements in network bandwidth, latency, and fault tolerance can be provided by careful selection of the protocols, choice of network topology, details of interconnection wiring, and basic wire driving technologies. We examine the improvements in some of these areas as part of the Transit project at the MIT Artificial Intelligence Laboratory View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An efficient algorithm for solving the token distribution problem on k-ary d-cube networks

    Publication Year: 1994, Page(s):175 - 182
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (416 KB)

    In parallel programs where the problem data is dynamically generated, it is very useful to be able to rely on an efficient load balancing algorithm. The token distribution problem (TDP) is a generalization of the static load balancing problem. The paper describes a novel algorithm for solving the TDP for k-ary d-cube topology networks. Compared to other algorithms, our method is more general and d... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Mathematic-physical engine: parallel processing for modeling and simulation of physical phenomena

    Publication Year: 1994, Page(s):366 - 373
    Cited by:  Papers (1)  |  Patents (6)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (344 KB)

    Mathematical modeling of various physical phenomena, ranging from fluid flow to electromagnetic field dynamics, from thermal patterns inside a semiconductor wafer to thermal patterns on a geophysical scale, and from combustion inside an automotive cylinder to ion transport across a biological membrane, is a challenge for the 1990s and beyond. These processes share in common a basic underlying stru... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A massively parallel implementation of pattern classifiers on SIMD and MIMD architectures

    Publication Year: 1994, Page(s):9 - 16
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (352 KB)

    Parallel multi-layer classifier architectures with an increasing hierarchical order have offered much flexibility in design to deal with a wide variety of properties. The model of pipeline processing is especially appropriate for realising such architectures. This has provided hierarchical classifiers a distinct advantage in real-time applications to cope with the important demand for high operati... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Building a better butterfly: the multiplexed metabutterfly

    Publication Year: 1994, Page(s):65 - 72
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (400 KB)

    Multistage networks are important in a wide variety of applications. Expander-based networks, such as multibutterflies, are a tremendous improvement over traditional butterflies in both fault and congestion tolerance. However, multibutterflies cost at least twice as much in chips and wiring as butterflies. It is also impossible to build large multibutterflies due to their wiring complexity. We sho... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Message-based efficient remote memory access on a highly parallel computer EM-X

    Publication Year: 1994, Page(s):135 - 142
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (448 KB)

    Communication latency is central to multiprocessor design. This report presents the design principles of EM-X multiprocessor towards tolerating communication latency. Multi-threading principle is built in the EM-X to overlap communication and computation for latency tolerance. In particular, we present two types of hardware support for remote memory access: (1) priority-based packet scheduling for... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Quicksort and permutation routing on the hypercube and de Bruijn networks

    Publication Year: 1994, Page(s):238 - 245
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (356 KB)

    We consider the problems of sorting and routing on some unbounded interconnection networks, namely hypercube and de Bruijn network. We first present two efficient implementations of quicksort on the hypercube. The first algorithm sorts N items on an N-node hypercube, one item per node, in O((log2 N)/(log log N)) time with high probability, while the other one sorts N items on an (N/log ... View full abstract»

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