检索结果-内蒙古大学图书馆

ieee symposium on On-Line Testing (IOLTS)

作者： Gilles Bizot Fabien Chaix Nacer-Eddine Zergainoh Michael Nicolaidis ARIS Group TIMA Laboratory-CNRS-INP UJF Grenoble France

the coming era of chips consisting of billions of gates foreshadows processors containing thousands of unreliable cores. In this context, high energy efficiency will be available, under the constraint that applications leverage the large amount of computing cores, while masking frequent faults of the chip. In this paper, an high-level method is proposed to map and manage a parallel application on an unreliable many-cores processor System on Chip. the approach takes into account versatile constraints relative to these processors (e.g. variability, core-level DVFS) and a generic algorithm is proposed. the distributed mapping process is based on the dynamic search of the best-suited processing node, upon task creation or node defect. An adaptive stop criteria is defined in order to balance the mapping impact and application efficiency gains. the validity of the proposition is assessed with high-level simulations, under different variability and application conditions.

关键词： Program processors Heuristic algorithms Equations Fault tolerance Fault tolerant systems System-on-chip Energy consumption

来源：评论

学校读者我要写书评

暂无评论

A performance evaluation of a temporally correct and optimistic concurrency control protocol

A performance evaluation of a temporally correct and optimis...

引用

international symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems (MASCOTS)

作者： A. Boukerche T.W. Tuck Sungbum Hong Department of Computer Science University of North Texas USA

In this paper we focus on the temporary return of data values that are incorrect for given transactional semantics and could have catastrophic effects similar to those in parallel and discrete event simulation. In many applications using on-line transaction processing (OLTP) environments, for instance, it is best to delay the response to a transaction's read request until it is either known or unlikely that a write message from an older update transaction will not make the response incorrect. Examples of such applications are those where aberrant behavior is too costly, and those in which precommitted data are visible to some reactive entity. In light of the avoidance of risk in this approach, we propose a risk-free multiversion temporally correct (RFMVTC) concurrency control algorithm. We discuss the algorithm, its implementation and report on the performance results of simulation models using a cluster of workstations.

关键词： Concurrency control Protocols Transaction databases distributed databases Relational databases Application software Clustering algorithms Database systems Computer science Discrete event simulation

来源：评论

学校读者我要写书评

暂无评论

Skyline Query Based on User Preference with MapReduce

Skyline Query Based on User Preference with MapReduce

引用

ieee international symposium on Dependable, Autonomic and Secure Computing (DASC)

作者： Yuanyuan Li Wenyu Qu Zhiyang Li Yujie Xu Changqing Ji Junfeng Wu College of Information Science and Technology Dalian Maritime University Dalian China College of software technology Dalian Jiaotong University Dalian China College of Physical Science and Technology Dalian University Dalian China Educational Technology and Computing Center Dalian Ocean University Dalian China

Skyline queries are useful in decision making applications. Skyline queries in highly mobile distributed environments have attracted many attentions recently due to the development of mobile internet device. the properties of distributed computing make skyline queries more complicated especially in any subspace. Conventional skyline algorithms do not support subspace skyline queries in distributed environment. In this paper, we focus on how to perform distributed skyline queries in any subspace according to user preference. So we propose a system model in a mobile and distributed environment. An efficient parallel algorithm for processing the Subspace Skyline Query (SSQ) using MapReduce is applied to the system model. this algorithm can report skyline points in any subspace. Meanwhile, a pruning strategy is also proposed in order to reduce the network communication and minimize the response time. We conduct experiments on real and synthetic data. Experimental results indicate that our SSQ algorithm is much more efficient. Furthermore, the pruning strategy can further improve the performance of the algorithm.

关键词： Indexes distributed databases Educational institutions Mobile communication Computational modeling Servers parallel algorithms

来源：评论

学校读者我要写书评

暂无评论

Experience report: A characteristic study on out of memory errors in distributed data-parallel applications

Experience report: A characteristic study on out of memory e...

引用

international symposium on Software Reliability Engineering (ISSRE)

作者： Lijie Xu Wensheng Dou Feng Zhu Chushu Gao Jie Liu Hua Zhong Jun Wei University of Chinese Academy of Sciences State Key Lab of Computer Science Chinese Academy of Sciences

Out of memory (OOM) errors occur frequently in data-intensive applications that run atop distributed data-parallel frameworks, such as MapReduce and Spark. In these applications, the memory space is shared by the framework and user code. Since the framework hides the details of distributed execution, it is challenging for users to pinpoint the root causes and fix these OOM errors. this paper presents a comprehensive characteristic study on 123 real-world OOM errors in Hadoop and Spark applications. Our major findings include: (1) 12% errors are caused by the large data buffered/cached in the framework, which indicates that it is hard for users to configure the right memory quota to balance the memory usage of the framework and user code. (2) 37% errors are caused by the unexpected large runtime data, such as large data partition, hotspot key, and large key/value record. (3) Most errors (64%) are caused by memory-consuming user code, which carelessly processes unexpected large data or generates large in-memory computing results. Among them, 13% errors are also caused by the unexpected large runtime data. (4) there are three common fix patterns (used in 34% errors), namely changing the memory/dataflow-related configurations, dividing runtime data, and optimizing user code logic. Our findings inspire us to propose potential solutions to avoid the OOM errors: (1) providing dynamic memory management mechanisms to balance the memory usage of the framework and user code at runtime; (2) providing users with memory+disk data structures, since accumulating large computing results in in-memory data structures is a common cause (15% errors).

关键词： Runtime Sparks Programming Fault tolerance Fault tolerant systems Memory management Data structures

来源：评论

学校读者我要写书评

暂无评论

Proceedings of the 16th international workshop on high-level parallel programming models and supportive environments

IEEE International Symposium on Parallel and Distributed Pro...

引用

ieee international symposium on parallel and distributed processing Workshops and Phd Forum 2011年 1139-1140页

作者： Hoefler, Torsten Eigenmann, Rudolf Gerndt, Michael Müller, Frank Rasmussen, Craig Schulz, Martin Alam, Sadaf Balaji, Pavan Barrett, Richard Bode, Brett Bosilca, George Bronevetsky, Greg Danalis, Anthony De Supinski, Bronis Ding, Chen Fahringer, thomas Ishikawa, Yutaka Knüpfer, Andreas Mohr, Bernd Scholz, Sven-Bodo Skjellum, Tony Snir, Marc Tillier, Fabian Larsson, Jesper Willcock, Jeremiah Wolf, Felix Blue Waters Directorate NCSA University of Illinois Urbana-Champaign United States Purdue University United States Technische Universität München Germany North Carolina State University United States Los Alamos National Laboratory United States Lawrence Livermore National Laboratory United States Swiss National Supercomputing Centre Switzerland Argonne National Laboratory United States Sandia National Laboratories United States National Center for Supercomputing Applications United States University of Tennessee Knoxville United States University of Rochester United States University of Innsbruck Austria University of Tokyo Japan Technische Universität Dresden Germany Forschungszentrum Jülich Germany University of Herfordshire United Kingdom University of Alabama Birmingham United States University of Illinois Urbana-Champaign United States Microsoft United States Träff University of Vienna Austria Indiana University United States German Research School for Simulation Sciences Germany

来源：评论

学校读者我要写书评

暂无评论

Elasticity in a Task-based Dataflow Runtime through Inter-node GPU Work Stealing

Elasticity in a Task-based Dataflow Runtime Through Inter-no...

引用

ieee/ACM international symposium on Cluster Computing and the Grid (CCGRID)

作者： Joseph John Josh Milthorpe National Computational Infrastructure Canberra Australia School of Computing Australian National University Canberra Australia Oak Ridge National Laboratory Oak Ridge TN USA

ISBN: (数字)9798350395662

ISBN: (纸本)9798350395679

Most contemporary HPC programming models assume an inelastic runtime in which the resources allocated to an application remain fixed throughout its execution. Conversely, elastic runtimes can expand and shrink resources based on availability and/or dynamic application requirements. In this paper, we implement elasticity for PaRSEC, a task-based dataflow runtime, using inter-node GPU work stealing. In addition to supporting elasticity, we demonstrate that inter-node GPU work stealing can enhance the performance of imbalanced applications by up to 45%.

关键词： Cloud computing Runtime High performance computing Graphics processing units distributed databases Elasticity Programming parallel processing Dynamic scheduling Resource management

来源：评论

学校读者我要写书评

暂无评论

Simulating Heterogeneous Models on Multi-Core Platforms using Julia's Computing Language parallel Potential

Simulating Heterogeneous Models on Multi-Core Platforms usin...

引用

international Workshop/symposium on distributed Simulation and Real-Time applications (DS-RT)

作者： Sergey Suslov Michael Schiek Markus Robens Christian Grewing Stefan van Waasen ZEA-2 - Electronic Systems Research Center Jülich Germany

ISBN: (数字)9781728173436

ISBN: (纸本)9781728173443

this paper addresses the questions of high-level system modelling using heterogeneous multi-tool modelling environment on parallel multi-core processing systems for simulation acceleration. the modelling technique has been applied for high-level validation of a high-precision Indoor Positioning System for Motion Analysis (IPS-MA) developed in the Central Institute Electronic Systems (ZEA-2) of the Research Center Juelich GmbH. the heterogeneous modelling environment has been built using an implementation-level model designed in Matlab Simulink, a verification model for describing the system environment using Modelica language, and Julia language for automatic generation of binding modelling environment and parallelizing the simulation of the overall model. the approach showed a good flexibility in system description and verification in the multi-instrument modelling environment and a good performance gain due to simulation parallelism.

关键词： Handheld computers Computational modeling Software packages Patents Sensors

来源：评论

学校读者我要写书评

暂无评论

GPU accelerated substructuring methods for sparse linear systems

GPU accelerated substructuring methods for sparse linear sys...

引用

第十五届分布式计算及其应用国际学术研讨会

作者： Abal-Kassim Cheik Ahamed Frédéric Magoulès CUDA Research Center & Applied Mathematics and Systems Laboratory CentraleSupélec Université Paris Saclay

ISBN: (纸本)9781509035946

In this paper,we present and analyze parallel substructuring methods based on conjugate gradient method,a iterative Krylov method,for solving sparse linear systems on *** experiments performed on a set of matrices coming from the finite element analysis of large scale engineering problems,show the efficiency and robustness of substructuring methods based on iterative Krylov method for solving sparse linear systems in a context of a hybrid multi-core-GPU.

关键词： Graphics processing units Computer architecture Instruction sets Linear systems Programming Central processing Unit Hardware

来源：评论

学校读者我要写书评

暂无评论

Performance analysis of a multithreaded pricing algorithm on Cilk

Performance analysis of a multithreaded pricing algorithm on...

引用

ieee international symposium on High Performance Computing Systems and applications (HPCS)

作者： R.K. thulasiram P. thulasiraman C. Adiele D. Bondarenko Department of Computer Science University of Manitoba Winnipeg Canada

In this paper, we develop a multithreaded algorithm for pricing simple options and implement it on a 8 node SMP machine using MIT's supercomputer programming language Cilk. the algorithm dynamically creates lots of threads to exploit parallelism and relies on the Cilk runtime system to distribute the computation load. We present both analytical and experimental results and our results explain how Cilk could be used effectively to exploit parallelism in the given problem. the analytical results show that our algorithm has a very high average parallelism and hence Cilk is the target paradigm to implement the algorithm. We conclude from our implementation results that the size of the threads, the number of threads created, the load balancer the cost of spawning a thread are parameters that must be considered while designing the algorithm on the Cilk platform.

关键词： Performance analysis Pricing Yarn parallel processing Algorithm design and analysis Supercomputers Computer languages Heuristic algorithms Concurrent computing distributed computing

来源：评论

学校读者我要写书评

暂无评论

Bringing metacomputing to Scilab

Bringing metacomputing to Scilab

引用

ieee international symposium on Computer Aided Control System Design (CACSD)

作者： F. Desprez E. Fleury C. Gomez S. Steer S. Ubeda INRIA-Rhône-Alpes-ReMaP Montbonnot Saint Martin France INRIA Lorraine-Résédas LORIA Vandoeuvre-les-Nancy France INRIA-Rocquencourt-Méta2 Le Chesnay France LIP-ReMaP URA 1398 du CNRS Lyon France

Scilab, developed at INRIA in the Meta2 project, is a software similar to Matlab that allows scientific applications to be developed using an interactive environment on a workstation running Unix or Windows. In this paper, we present the current version of Scilab/sub ///, a extension of this software that supports the execution of parallel jobs within Scilab. Using simple commands and the same interface as Scilab, Scilab/sub /// allows users to start Scilab on a parallel machine or a network of workstations, therefore giving access to the computation power and huge aggregate memory sizes. From the Scilab interactive environment, users can spawn several other Scilab processes, executing remote Scilab scripts, which can then communicate between each other. Finally, processes can print/plot their results. there are many research directions around this project and we give an overview of the current and future developments.

关键词： Metacomputing Workstations Concurrent computing Sparse matrices parallel machines Computer interfaces Computer networks Signal processing algorithms distributed computing Application software

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：