检索结果-内蒙古大学图书馆

International conference on parallel and distributed Systems (ICPADS)

作者： Sook-Yeon Kim Kyung-Yong Chwa Department Computer Science KAIST Taejon South Korea

ISBN: (纸本)0818682272

A hypermesh, a versatile parallel architecture, is obtained from a 2-dimensional mesh by replacing each linear connection with a hyper-edge. We optimally embed multiple graphs into a hypermesh by a labeling strategy. This optimal embedding provides an optimal expansion, dilation and congestion at the same time. First, we label on an N-node graph G, possibly disconnected, such that this labeling makes it possible to optimally embed multiple copies of G into an N'/spl times/N' hypermesh when N' is divisible by N. Second, we show that many important classes of graphs have this labeling: for example, tree, cycle, mesh of trees and product graphs including mesh, torus, and hypercube. Third, we generalize these results to optimally embed multiple graphs into a multidimensional and possibly non-square hypermesh. This labeling strategy is applicable to the embeddings of other classes of graphs into a hypermesh.

关键词： Labeling Computer science parallel architectures Tree graphs parallel algorithms Hypercubes Ear Multidimensional systems Computer architecture Concurrent computing

来源：评论

学校读者我要写书评

暂无评论

Using cache optimizing compiler for managing software cache on distributed shared memory system

Proceedings of the Conference on High Performance Computing ...

引用

proceedings of the conference on High Performance computing on the Information Superhighway, HPC Asia'97 1997年 312-318页

作者： Nanri, Takeshi Sato, Hiroyuki Shimasaki, Masaaki Kyushu Univ Fukuoka Japan

On distributed shared memory (DSM) system, the optimization of memory access is very important for achieving good performance. In this paper we propose an optimizing compiler which controls a software cache system implemented on DSM. The software cache consists of static part related with compiler and dynamic part related with cache-managing runtime routines. Our compiler controls static part of the software cache by using information from static analysis. For applications whose behavior can only be dynamically determined, the compiler uses dynamic part of the software cache. We also propose application of RISC-oriented optimization techniques to parallel applications on our software cache system. We evaluate the efficiency of our compiler and RISC-oriented optimization techniques on distributed parallel machine, CM-5. Our results show that the compiler and the optimizations considerably improve the performance of basic linear algebra routines: matrix multiply, Cholesky decomposition and Gaussian elimination.

关键词： Program compilers

来源：评论

学校读者我要写书评

暂无评论

Information flow in a purpose-oriented access control model

Information flow in a purpose-oriented access control model

引用

International conference on parallel and distributed Systems (ICPADS)

作者： M. Yasuda T. Tachikawa K. Takizawa Department of Computers and Systems Engineering Tokyo Denki University Japan

In distributed applications, a group of multiple objects are cooperating to achieve some objectives. An object is modeled as a pair of data structures and operations. Each object is manipulated through an operation supported by the object and then the operation may further invoke operations of other objects, i.e., nested operations. The purpose-oriented access rules indicate which operation on each object can invoke operations of other objects. The information flow among the objects occur if the requests and responses of the operations carry some data. Only the purpose-oriented access rules which imply the legal information flow are allowed. We discuss how to test the access rules if the information flow occurring in the nested invocation of the operations is legal.

关键词： Access control Law Legal factors Information security Data structures distributed computing Systems engineering and theory Application software System testing Client-server systems

来源：评论

学校读者我要写书评

暂无评论

Fast reductions on a network of workstations

Fast reductions on a network of workstations

引用

International conference on High Performance computing

作者： S. Srinivasan M. Lyell J. Wehrwein P.F. Reynolds Mystech Associates Inc. Falls Church VA USA Department of Computer Science University of Virginia Charlottesville VA USA

Reduction operations are very useful in parallel and distributed computing, with applications in barrier synchronization, distributed snapshots, termination detection, global virtual time computation, etc. In the context of parallel discrete event simulations, we have previously introduced a class of adaptive synchronization algorithms based on fast reductions. We explore the implementation of fast reductions on a popular high performance computing platform-a network of workstations. The specific platform is a set of Pentium Pro PC's running the Linux operating system interconnected by Myrinet-a Gbps network. The general reduction model on which our synchronization algorithms are based is introduced first, followed by a description of how this model can be implemented. We discuss several design trade offs that must be made in order to achieve the driving goal of high speed reductions and provide innovative algorithms to meet the correctness and performance requirements of the reduction model.

关键词： Workstations Discrete event simulation distributed computing Concurrent computing parallel programming Hardware Computer science Test facilities Springs Contracts

来源：评论

学校读者我要写书评

暂无评论

Performance of the r-truncated Benes networks under randomized routing algorithms

Performance of the r-truncated Benes networks under randomiz...

引用

International conference on parallel and distributed Systems (ICPADS)

作者： H. El-Sayed A. Youssef Department of Electrical Engineering and Computer Science George Washington University Washington D.C. DC USA

Benes networks have the potential for balanced traffic, fewer conflicts, and can route any permutation in one pass due to their multiplicity of paths. Omega networks, on the other hand, have fast set-up and low hardware cost, but could take more than one pass to route a permutation. This paper introduces a new class of networks referred to as r-truncated Benes, which is a Benes network with r randomization stages eliminated. Using randomized routing, we will show that r-truncated Benes networks is an excellent trade-off between Omega and Benes networks. In particular, it will be shown that the I-truncated Benes network out performs Omega and is also superior to Benes and other truncated Benes networks in cost and performance.

关键词： Routing Costs Hardware Multiprocessor interconnection networks parallel algorithms Telecommunication traffic Ambient intelligence parallel processing Computer networks Concurrent computing

来源：评论

学校读者我要写书评

暂无评论

ELMO: extending (sequential) languages with migratable objects-compiler support

ELMO: extending (sequential) languages with migratable objec...

引用

International conference on High Performance computing

作者： R.J. Richards B. Ramkumar S.G. Rathnam Department of Electrical and Computer Engineering University of Iowa Iowa USA

Efficient task migration is an important feature in parallel and distributed programs, in particular to support checkpointing and recovery for fault tolerance. It is also very useful in distributed environments like networks of workstations where external loads are often unpredictable and dynamic in nature. We propose simple language extensions (ELMO) to existing sequential programming languages like C, Fortran or C++, that provide an object based task parallel execution model. Tasks may be dynamically created, are location transparent, and may be migrated or check pointed transparently by the system. ELMO's language features, together with the requisite compiler support is presented.

关键词： Checkpointing Computer networks Concurrent computing Debugging Fault tolerance Workstations Delay parallel programming Runtime Program processors

来源：评论

学校读者我要写书评

暂无评论

Molecular mechanical calculations on parallel and distributed processors

Molecular mechanical calculations on parallel and distribute...

引用

International conference on High Performance computing in the Asia-Pacific Region

作者： Youngdo Won Department of Chemistry Hanyang University Seoul South Korea

The molecular mechanical potential function is widely used in molecular modeling and simulation research. The most CPU time consuming parts of molecular mechanical potential are the nonbonding interaction terms. An efficient parallel algorithm for nonbonding energy calculation is outlined and its implementation is tested on a variety of parallel and distributed processing elements. As minimal parallel constructs are added, the current implementation does not modify nor slow down the serial algorithm. Load balancing, flexible enough to accommodate local loads of each PE and optimization of list updating procedure are desired and under development.

关键词： Atomic measurements Biological system modeling Bonding Mechanical energy Chemistry parallel algorithms Testing distributed processing Load management Microprocessors

来源：评论

学校读者我要写书评

暂无评论

High performance computing for direct turbulence simulation in computational fluid dynamics

High performance computing for direct turbulence simulation ...

引用

International conference on High Performance computing in the Asia-Pacific Region

作者： G. Coussement B. Troff J. Ryan H. Yamazaki M. Yoshida T. Iwamiya Office National d'Etudes et de Recherches Aérospatrales (ONERA) Chatillon France National Aerospace Laboratory Chofu Tokyo Japan

The remarkable advances obtained in prediction and design in computational fluid dynamics (CFD) which requires scalable computing power in terms of speed and memory can currently be sustained with parallel supercomputers. This paper presents cooperative work between ONERA (France) and NAL (Japan) to investigate the potential of distributed environments for high performance computing in CFD. For this purpose a Navier-Stokes flow code was integrated into a distributed processor environment with two parallelization strategies: one based on domain partitioning with message passing; and another one based on loop-parallelization with compiler directives. We present in this paper the results obtained on different computers such as CRAY, PARAGON and NAL NWT supercomputers.

关键词： High performance computing Computational modeling Computational fluid dynamics Supercomputers Collaborative work Message passing Navier-Stokes equations Finite element methods Concurrent computing Protocols

来源：评论

学校读者我要写书评

暂无评论

Predictable network computing

Predictable network computing

引用

International conference on distributed computing Systems

作者： A. Polze M. Werner G. Fohler Software Engineering Institute CMU USA Humboldt University of Berlin Berlin Germany

Clusters of networked commercial, off the shelf (COTS) workstations are presently used for computation intensive tasks that were typically assigned to parallel computers in the past. However, it is hardly possible to predict the timing behavior of such systems or to give guarantees about execution times. We show how our SONiC (Shared Objects Net-interconnected Computer) system can control timing and partitioning of a workstation as a step towards a distributed real time system built from COTS components. SONiC provides a class based programming interface for creation of replicated shared objects of arbitrary, user defined sizes. Weak consistency protocols are employed to improve system performance. Our scheduling service ensures the requested interactive behavior of a workstation while simultaneously giving a specified number of CPU cycles to parallel tasks. Using offline scheduling methods we are able to implement real time guaranteed services on COTS workstations.

关键词： Computer networks Workstations Concurrent computing Timing distributed computing Time sharing computer systems Control systems Real time systems Protocols System performance

来源：评论

学校读者我要写书评

暂无评论

Reducing overhead in implementing fine-grain parallel data-structures of a dataflow language on off-the-shelf distributed-memory parallel computers

Reducing overhead in implementing fine-grain parallel data-s...

引用

Annual Hawaii International conference on System Sciences (HICSS)

作者： S. Kusakabe T. Nagai K. Inenaga M. Amamiya Kyushu University Japan

In order to show the feasibility of a fine-grain dataflow computation scheme, we are implementing a fine-grain dataflow language on off-the-shelf computers, using a fine-grain multithread approach. Fine-grain parallel data-structures such as I-structures provide high level abstraction to easily write programs with potentially high parallelism. The results of preliminary experiments on a distributed memory parallel machine indicate that the performance inefficiency related to fine-grain parallel data-structures in the naive implementation is mainly caused by the calculation of the local address for distributed data, and the frequent fine-grain data access using message passing. In order to reduce the addressing overhead, we introduce a two-level table addressing technique. We employ a caching mechanism and a grouping mechanism for the fine-grain data access. The preliminary performance evaluation results indicate that these techniques are effective to improve the performance.

关键词： Automatic control Runtime Message passing Concurrent computing distributed computing parallel processing parallel programming Timing Hardware parallel machines

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：