By Topic

Parallel and Distributed Computing, Applications and Technologies, 2008. PDCAT 2008. Ninth International Conference on

Date 1-4 Dec. 2008

Filter Results

Displaying Results 1 - 25 of 90
  • [Front cover]

    Publication Year: 2008 , Page(s): C1
    Save to Project icon | Request Permissions | PDF file iconPDF (7129 KB)  
    Freely Available from IEEE
  • [Title page i]

    Publication Year: 2008 , Page(s): i
    Save to Project icon | Request Permissions | PDF file iconPDF (48 KB)  
    Freely Available from IEEE
  • [Title page iii]

    Publication Year: 2008 , Page(s): iii
    Save to Project icon | Request Permissions | PDF file iconPDF (86 KB)  
    Freely Available from IEEE
  • [Copyright notice]

    Publication Year: 2008 , Page(s): iv
    Save to Project icon | Request Permissions | PDF file iconPDF (46 KB)  
    Freely Available from IEEE
  • Table of contents

    Publication Year: 2008 , Page(s): v - x
    Save to Project icon | Request Permissions | PDF file iconPDF (148 KB)  
    Freely Available from IEEE
  • Message from General Chairs

    Publication Year: 2008 , Page(s): xi
    Save to Project icon | Request Permissions | PDF file iconPDF (72 KB) |  | HTML iconHTML  
    Freely Available from IEEE
  • Message from Programme Committee Chairs

    Publication Year: 2008 , Page(s): xii
    Save to Project icon | Request Permissions | PDF file iconPDF (73 KB) |  | HTML iconHTML  
    Freely Available from IEEE
  • Conference organization

    Publication Year: 2008 , Page(s): xiii
    Save to Project icon | Request Permissions | PDF file iconPDF (87 KB)  
    Freely Available from IEEE
  • Program Committee

    Publication Year: 2008 , Page(s): xiv - xv
    Save to Project icon | Request Permissions | PDF file iconPDF (92 KB)  
    Freely Available from IEEE
  • list-reviewer

    Publication Year: 2008 , Page(s): xvi
    Save to Project icon | Request Permissions | PDF file iconPDF (87 KB)  
    Freely Available from IEEE
  • Looking toward Exascale Computing

    Publication Year: 2008 , Page(s): 3
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (117 KB) |  | HTML iconHTML  

    Recently, Argonne National Laboratory installed a half petaflop Blue Gene/P system. It is the world's fastest open science supercomputer. With 163,840 cores, the machine is beginning to provide insight on how we might build future platforms as we scale toward exascale computing. There are many challenges, including the dramatic shift to multicore, the cost of electric power, and the need for robust fault management. In this talk I will focus on the architecture and system software challenges we face as we continue to attack ever-larger computational problems. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Virtual Organizations By the Rules

    Publication Year: 2008 , Page(s): 4
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (121 KB) |  | HTML iconHTML  

    Increasingly, collaborative activities in science are using the concept of virtual organization as an organizing principle. One benefit of viewing these collaborations from an organizational perspective is that there is a long history of studying how organizations can be structured to function effectively. Many of these organizational principles have been reflected in the design of enterprise architectures and the use of service oriented architecture concepts as an implementation vehicle for capturing these organizational constructs. One approach to meeting organizational requirements in systems architecture has been to express organizational structure in terms of business roles, business processes and business rules. To date however, this type of analysis and associated infrastructure tools has not been applied in any consistent way to the concept of virtual organizations and their associated scientific applications. In this talk, the author explores these established approaches to business IT systems and their applicability to the virtual organizations that are being created to support scientific endeavors. As an example, I will describes how data management policies for virtual organization can be expressed as business rules, and implemented via existing business rules engines. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Case Studies in Computer Network Measurement

    Publication Year: 2008 , Page(s): 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (39 KB) |  | HTML iconHTML  

    A network lies at the heart of many modern computer systems including most distributed and parallel systems. A good understanding of the behaviour and performance of the network is, therefore, a prerequisite for the understanding the behaviour and performance of many computer systems. Understanding network behaviour and performance is challenging because networks are among mankind's largest and most complex creations and because most networks, including the Internet, are not designed with network measurement in mind. As a consequence, much network measurement relies on inference from data that does not directly describe the item of interest. Measurement also often requires a widely spread physical infrastructure, for example a network of monitoring machines. This is expensive both in terms of the hardware required and the resources needed to deploy and maintain the systems. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Layered Virtual Organization Architecture for Grid

    Publication Year: 2008 , Page(s): 9 - 16
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (439 KB) |  | HTML iconHTML  

    Virtual organizations (VO) are widely accepted in the grid and other distributed computing environments. However, there are few effective VO implementations. This paper presents a layered architecture to construct Agora, an implementation of VO. Agora manages users, resources, and agora instances, provides policies to support a DAC/MAC-hybrid cross-domain access control mechanism, and maintains the context of operations. The Agora architecture consists of three layers. At the bottom is the physical layer containing external resources, then an abstraction RController is introduced to manipulate external resources. Above the physical layer, all the involved entities, including users, resources, and agoras, are abstracted as GNodes, and a naming layer is introduced to manage these GNodes. At the top, the logic layer implements all the Agora functionalities. This architecture has been implemented in Vega GOS and applied in the China National Grid and other grid platforms. The evaluation shows that the architecture provides minimal but sufficient VO functionalities while keeping decentralization, flexibility, simplicity, and effectiveness. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Operating System-Level Virtual Organization Support in XtreemOS

    Publication Year: 2008 , Page(s): 17 - 24
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (302 KB) |  | HTML iconHTML  

    In computational grids, a virtual organization (VO) is a dynamic coupling of multiple Linux/Unix nodes for resource sharing under specific polices. Currently, VO support functionalities are generally implemented as grid middleware. However, the usability of grids is often impaired by the complexity of configuring and maintaining a new layer of security infrastructure as well as adapting to new interfaces of security enabled services. In this paper, we present an OS-level approach to provide native VO support functionalities, which is a part of XtreemOS project [18]. Our approach adopts pluggable frameworks existing in current OS as extension points to implement VO support, avoiding modification of kernel codes and easily turning traditional OSes into grid-aware ones. The performance evaluation of NAS parallel benchmarks (NPB) shows that our current implementation incurs trivial overhead on original systems. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Scalable Contract Net Based Resource Allocation Strategies for Grids

    Publication Year: 2008 , Page(s): 25 - 32
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (345 KB) |  | HTML iconHTML  

    Different economic models have been applied to the resource allocation problem in grids. Contract net is one such model in which a job is announced to the resources, the resources bid for the job, and then the broker chooses one bid from the received bids based on some policy. The job is awarded to the resource with the winning bid. In this paper, we first propose a contract net based resource allocation policy for grids based on an earlier policy, that performs better than other existing policies. We next perform a scalability analysis of the policy to show that it is not scalable. Finally we present strategies to make contract net scalable in large grid resource allocation problems. We present detailed simulation results to evaluate the performances of the proposed strategies. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An Experimental Analysis for Memory Usage of GOS Core

    Publication Year: 2008 , Page(s): 33 - 36
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (317 KB) |  | HTML iconHTML  

    As grid software and grid applications grow in size and complexity, memory usage experiment and analysis have become more and more essential. In this paper, we investigate an antitestcase-based approach to analyze the memory usage of GOS core software, and present a set of anti-testcases, which describe predefined memory problems of GOS core. This approach is applied with a set of automated test tools in controlled experiments including basic memory requirements, memory errors and memory usage of GOS core. With the experiment and analysis, we find several memory usage results of GOS core and they are valuable for other systems based on JVM, Tomcat and Axis. We also give some JVM memory configuration suggestions for better performance in the production server. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Data-Parallel Algorithm to Reliably Solve Systems of Nonlinear Equations

    Publication Year: 2008 , Page(s): 39 - 46
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (297 KB) |  | HTML iconHTML  

    Numerical methods based on interval arithmetic are efficient means to reliably solve nonlinear systems of equations. Algorithm bc3revise is an interval method that tightens variables' domains by enforcing a property called box consistency. It has been successfully used on difficult problems whose solving eluded traditional numerical methods.We present a new algorithm to enforce box consistency that is simpler than bc3revise, faster, and easily data parallelizable. A parallel implementation with Intel SSE2 SIMD instructions shows that an increase in performance of up to an order of magnitude and more is achievable. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A New Iterative Eliiptic PDE Solver on a Distributed PC Cluster

    Publication Year: 2008 , Page(s): 47 - 53
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (305 KB)  

    Advancement in parallel computers technology has greatly influenced the numerical methods used for solving partial differential equations (pdes). A lot of attention has been devoted to the development of numerical schemes which are suitable for the parallel environment. In this work, we investigate the parallel implementation of the four-point Modified Explicit Decoupled Group (MEDG) method which was introduced by Ali and Ng (2007) as a fast solver for the two dimensional Poisson pde. The method was shown to be more superior than all the methods belonging to the four-points explicit group family namely the Explicit Group (EG) [8], Explicit Decoupled Group (EDG) [1] and Modified Explicit Group (MEG) [7]. This paper presents the preliminary results of the parallel algorithms implemented on a distributed memory PC cluster. Two parallelizing strategies comprising of the two-color zebra and the four-color chessboard orderings in solving a two dimensional Poisson model problem will be discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An Effective Structure for Algorithmic Design and a Parallel Prefix Algorithm on Metacube

    Publication Year: 2008 , Page(s): 54 - 61
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (206 KB) |  | HTML iconHTML  

    Metacube is an attractive, hypercube-like interconnection network that can connect an extremely large number of nodes with a small node degree while keeping a relatively short diamemter. A Metacube MC(k,m) connects 22 k m+k nodes with only k+m links per node. Metacube can be used to build parallel computing systems of very large scale with a small number of links per node. In this paper, we propose a new presentation of Metacube for algorithmic design on Metacube. Based on the new presentation, we give an efficient algorithm for parallel prefix computation on Metacubes that runs in 2km(k+1)+k communication steps and 2k+1m+2k computation steps. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Parallel Algorithm for Block Tridiagonal Systems

    Publication Year: 2008 , Page(s): 62 - 65
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (160 KB) |  | HTML iconHTML  

    A parallel algorithm, namely parallel block diagonal dominant (PBDD) algorithm, is proposed to solve block tridiagonal linear systems on multi-computers. This algorithm is based on divided-and-conquer idea of the PDD method. When the systems is strictly block diagonal dominant, the PBDD is highly parallel and provides approximate solutions that equals to the exact solutions within machine accuracy. The PBDD method has been implemented on a 64-node multi-computer. The analytic results match closely with the results measured from the numerical experiments. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An Efficient Parallel Algorithm for H.264/AVC Encoder

    Publication Year: 2008 , Page(s): 66 - 69
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (249 KB) |  | HTML iconHTML  

    In this paper, a bit-rate estimation method is proposed, which could improve the performance of the H.264 encoder and avoid the strict data-dependencies of mode decision. After that, an efficient parallel algorithm for H.264 encoder with CABAC entropy coding is presented based on the Macro-Block Region Partition (MBRP) parallel method and the bit-rate estimation technique. Simulation results show that, the proposed parallel algorithm could improve the performance of H.264 encoder efficiently while maintaining the similar RD performance as JM 10.2. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parallelization and Acceleration Scheme of Multilevel Fast Multipole Method

    Publication Year: 2008 , Page(s): 70 - 73
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (287 KB) |  | HTML iconHTML  

    The iterative methods such as BiCGStab for solving electromagnetic field integer equations have a complexity of O(N2), which can be reduced to O(N logN) by multilevel fast multipole method (MLFMM). For large scale problems, MLFMM should be parallelized, and the iterative convergence can be accelerated by preconditioners such as incomplete inverse triangular factorization preconditioner. The interpolation based on spherical harmonic transform at each level of MLFMMpsilas octree can be further accelerated by FFT. Based on this acceleration scheme tested on distributed cluster, the results show this algorithm is feasible. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parallel Approximate Multi-Pattern Matching on Heterogeneous Cluster Systems

    Publication Year: 2008 , Page(s): 74 - 79
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (518 KB) |  | HTML iconHTML  

    For the given multiple patterns and a text string, firstly, a perfect hash function is constructed, the patterns are transformed into the unique pairs of integer values in parallel by the perfect hash function, the corresponding integer values are stored in a global hash table, and a recursion expression for computing hash function value of the signatures of each sub-string of text is also proposed. Secondly, based on divisible load principle, a linear programming model for the optimal text distribution strategy is created and a parallel approximate multi-pattern matching algorithm allowing one error is presented on the heterogeneous cluster system which processors have different computing speeds and distinct communication capabilities and different memory sizes by taking into account computation and communication startup time and using the assigned processor distribution order. The experimental results on the cluster system of heterogeneous personal computers show that the presented parallel algorithm is averagely 25% faster than that one using the even text distribution strategy, and it obtains a nearly linear speedup and good scalability. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Switch-Based Packing Technique for Improving Token Coherence Scalability

    Publication Year: 2008 , Page(s): 83 - 90
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (172 KB) |  | HTML iconHTML  

    Traditional cache coherence protocols either provide low latency cache misses (snooping protocols) or bandwidth efficiency (directory protocols). To simultaneously capture the best attributes of traditional protocols, Token Coherence has been recently proposed. This protocol can quickly resolve cache misses by transient requests. However, since transient requests are unordered messages, they may sometimes fail in solving cache misses mainly due to the occurrence of protocol races. Thus, when the completion of cache misses is not possible by transient requests, Token Coherence uses a starvation prevention mechanism to ensure their completion. Although several implementation options of starvation prevention mechanisms have been proposed, all of them are broadcast-based. This fact represents a large detriment to the Token Coherence scalability. To tackle this problem, in this work we apply a switch-based packing technique that alleviates the harm of broadcast messages and improves the protocol scalability. View full abstract»

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