检索结果-内蒙古大学图书馆

Annual international symposium on Computer Architecture, ISCA

作者： P. Stenstrom T. Joe A. Gupta Department of Computer Engineering Lund University Lund Sweden Computer Systems Laboratory University of Stanford CA USA

Two interesting variations of large-scale shared-memory ma- chines that have recently emerged are cache-coherent non- uniform-memory-access machines (CC-NUMA) and cache- only memory architectures (COMA). They both have distributed main memory and use directory-based cache coherence. Unlike CC-NUMA, however, COMA machines automatically migrate and replicate data at the main-memory level in cache-line sized chunks. This paper compares the performance of these two classes of machines. We first present a qualitative model that shows that the relative performance is primarily determined by two factors: the relative magnitude of capacity misses versus coherence misses, and the granularity of data partitions in the application. We then present quantitative results using simulation studies for eight parallel applications (including all six applications from the SPLASH benchmark suite). We show that COMA's potential for performance improvement is limited to applications where data accesses by different processors are finely interleaved in memory space and, in addition, where capacity misses dominate over coherence misses. In other situations, for example where coherence misses dominate, COMA can actually perform worse than CC-NUMA due to increased miss latencies caused by its hierarchical directories. Finally, we propose a new architectural alternative, called COMA-F, that combines the advantages of both CC-NUMA and COMA.

关键词： Delay Permission Computer architecture Large-scale systems Memory architecture distributed decision making Multiprocessor interconnection networks Bandwidth Broadcasting Laboratories

来源：评论

学校读者我要写书评

暂无评论

The Cactus Computational Toolkit and Using distributed Computing to Collide Neutron Stars 99

The Cactus Computational Toolkit and Using Distributed Compu...

引用

proceedings of the 8th IEEE international symposium on High Performance distributed Computing

作者： Gabrielle Allen Tom Goodale Joan Massó Edward Seidel Max-Planck-Institut-für-Gravitationsphysik

ISBN: (纸本)9780769502878

We are developing a system for collaborative research and development for a distributed group of researchers at different institutions around the world. In a new paradigm for collaborative computational science, the computer code and supporting infrastructure itself becomes the collaborating instrument, just as an accelerator becomes the collaborating tool for large numbers of distributed researchers in particle physics. The design of this "Collaboratory" allows many users, with very different areas of expertise, to work coherently together, on distributed computers around the world. Different supercomputers may be used separately, or for problems exceeding the capacity of any single system, multiple supercomputers may be networked together through high speed gigabit networks. Central to this Collaboratory is a new type of community simulation code, called "Cactus". The scientific driving force behind this project is the simulation of Einstein's equations for studying black holes, gravitational waves, and neutron stars, which has brought together researchers in very different fields from many groups around the world to make advances in the study of relativity and astrophysics. But the system is also being developed to provide scientists and engineers, without expert knowledge of parallel or distributed computing, mesh refinement, and so on, with a simple framework for solving any system of partial differential equations on many parallel computer systems, from traditional supercomputers to networks of workstations.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Data warehouse middleware towards an unified resource management

Data warehouse middleware towards an unified resource manage...

引用

IFIP/IEEE international symposium on Integrated Network Management

作者： P. Assis J.A. Martins DEE ISEP Instituto Politecnico do Porto Porto Salvo Portugal DET Universidade de Aveiro Aveiro Portugal

ISBN: (纸本)0780367197

The global quest for information is driving the communication infrastructure development, demanding for non-stop systems availability with higher capabilities. Hence, the management improvement is a trivial requirement! The ability to integrate in a unique platform all types of manageable resources, e.g. an application, an implementation of a service, a device, and so on, is in the aim of new management approaches. This work presents the concepts for a data warehouse (DW) middleware layer, which will enable the interface between management applications and the manageable resources, regardless of the supported management protocols. The goals we intent to attain are: achieve the integration of existing management domains, and the usage of DW facilities in the establishment of proactive management.

关键词： Data warehouses Middleware Resource management Protocols Availability Quality management distributed databases Automation Database languages Data mining

来源：评论

学校读者我要写书评

暂无评论

Scrambled storage for parallel memory systems 88

引用

proceedings of the 15th Annual international symposium on Computer architecture

作者： D. Lee York Univ. North York Ont. Canada

ISBN: (纸本)9780818608612

A scrambled storage scheme is proposed for storing arrays of NXN elements in N = 2n parallel memory modules to allow conflict-free access to various array partitions. It is shown that the scheme allows conflict-free access to rows, columns, square blocks, and distributed blocks of stored arrays. An alternative way of achieving the desired accessibility would use Budnik and Kuck's nonuniform skewed storage [3]; in this case, addressing hardware would require Ο(nX2n) exclusive-or circuits. The proposed scheme has, however, the advantage of simplifying address generation; addressing hardware requires n exclusive-or circuits only. Some of the important questions of scrambling/unscrambling data through a proposed interconnection network are discussed.

关键词：

来源：评论

学校读者我要写书评

暂无评论

DustDoctor: A self-healing sensor data collection system

DustDoctor: A self-healing sensor data collection system

引用

international symposium on Information Processing in Sensor Networks (IPSN)

作者： Mohammad Maifi Hasan Khan Hossein Ahmadi Gulustan Dogan Kannan Govindan Raghu Ganti Theodore Brown Jiawei Han Prasant Mohapatra Tarek Abdelzaher University of Illinois USA University of California Davis USA IBM Thomas J. Watson Research Center USA City University of New York USA

This demonstration presents a tool, called DustDoctor, for troubleshooting sensor data fusion systems in which data is combined from multiple heterogeneous sources to compute actionable information. Application examples include target detection, critical infrastructure monitoring, and participatory sensing. In such systems, the correctness of end results may become compromised for a variety of possible reasons, such as node malfunction, bugs, environmental conditions unfavorable to certain sensors, or assumption mismatches (such as use of incompatible units on different nodes of the same distributed computation). DustDoctor adapts algorithms borrowed from previous discriminative mining literature to analyze data fusion flow graphs, called provenance graphs, and isolate sources and conditions correlated with anomalous results. This information is subsequently used to isolate malfunctioning components or filter out erroneous reports. We demonstrate our approach on MicaZ motes, running a simple data collection application, where users are allowed to inject a variety of different emulated faults, leaving it to DustDoctor to find and isolate them to prevent contamination of fusion results.

关键词： Labeling Data mining Computer bugs Context distributed databases Government Pollution measurement

来源：评论

学校读者我要写书评

暂无评论

Slipstream execution mode for CMP-based multiprocessors

Slipstream execution mode for CMP-based multiprocessors

引用

IEEE symposium on High-Performance Computer Architecture

作者： K.Z. Ibrahim G.T. Byrd E. Rotenberg Department of Electrical and Computer Engineering North Carolina State University USA

Scalability of applications on distributed shared-memory (DSM) multiprocessors is limited by communication overheads. At some point, using more processors to increase parallelism yields diminishing returns or even degrades performance. When increasing concurrency is futile, we propose an additional mode of execution, called slipstream mode, that instead enlists extra processors to assist parallel tasks by reducing perceived overheads. We consider DSM multiprocessors built from dual-processor chip multiprocessor (CMP) nodes with shared L2 cache. A task is allocated on one processor of each CMP node. The other processor of each node executes a reduced version of the same task. The reduced version skips shared-memory stores and synchronization, running ahead of the true task. Even with the skipped operations, the reduced task makes accurate forward progress and generates an accurate reference stream, because branches and addresses depend primarily on private data. Slipstream execution mode yields two benefits. First, the reduced task prefetches data on behalf of the true task. Second, reduced tasks provide a detailed picture of future reference behavior, enabling a number of optimizations aimed at accelerating coherence events, e.g., self-invalidation. For multiprocessor systems with up to 16 CMP nodes, slipstream mode outperforms running one or two conventional tasks per CMP in 7 out of 9 parallel scientific benchmarks. Slipstream mode is 12-19% faster with prefetching only and up to 29% faster with self-invalidation enabled.

关键词： Concurrent computing parallel processing Scalability Degradation Prefetching Performance gain Acceleration Coherence Multiprocessing systems Character generation

来源：评论

学校读者我要写书评

暂无评论

Notice of Violation of IEEE Publication Principles: Multi-paradigm and Multi-grain parallel Execution Model Based on SMP-Cluster

Notice of Violation of IEEE Publication Principles: Multi-pa...

引用

IEEE John Vincent Atanasoff international symposium on Modern Computing, JVA

作者： Jue Wang Changjun Hu Jianxin Lai Yudi Zhao Suqin Zhang Computer Science and Technology Department Tsinghua University Beijing China Inf. Eng. Sch. Univ. of Sci. & Technol. Beijing Information Engineering School University of Science and Technology Beijing China Information Engineering School University of Science and Technology Beijing Information Engineering School University of Science and Technology Beijing China

Notice of Violation of IEEE Publication Principles "Multi-paradigm and Multi-grain parallel Execution Model Based on SMP-Cluster," by Jue Wang, Changjun Hu, Jianxin Lai, Yudi Zhao, and Suqin Zhang, in the proceedings of the IEEE John Vincent Atanasoff 2006 international symposium on Modern Computing, 2006 (JVA '06) pp.266-272 After careful and considered review of the content and authorship of this paper by a duly constituted expert committee, this paper has been found to be in violation of IEEE's Publication Principles. This paper contains portions of original text from the papers cited below. The original text was copied without appropriate attribution. "Comparing the OpenMP, MPI, and Hybrid Programming Paradigm on an SMP Cluster," by Gabriele Jost, Haoqiang Jin, Dieter an Mey, and Ferhat F. Hatay, in IEEE Transactions on parallel and distributed systems, 10(2), February 1999 In this paper, we propose a multi-paradigm and multi-grain parallel execution model based on SMP-Cluster, which integrates coarse grain, mid grain and fine grain parallelism. Multiple paradigms supported by our model include task parallel, data parallel, sequential execution, data pipeline and task-farming paradigm. It can be achieved by extending the OpenMP specification, and the extensions include directives for computing resource partition, data distribution and alignment, sequential execution and data pipeline, and functions for Master/Slave model in Macro-Task group. We also compare the performance of different implementations of three benchmark applications, using the same numerical algorithm but employing different programming approaches

关键词：

来源：评论

学校读者我要写书评

暂无评论

Enhancing NWS for use in an SNMP managed Internetwork

Proceedings of the International Parallel Processing Symposi...

引用

proceedings of the international parallel Processing symposium, IPPS 2000年 506-511页

作者： Busby Jr., Robert E. Neilsen, Mitchell L. Andresen, Daniel AT&T Network Operations Bridgeton United States

The Network Weather Service (NWS) is a distributed resource monitoring and utilization prediction system, employed as an aid to scheduling jobs in a metacomputing environment. We have enhanced the Network Weather Service by providing a broader metric set, forecast trend visualization, and asynchronous notification. Three new components have been added to the NWS system. The first of these is an SNMP Sensor that may be used to collect data being tracked by SNMP agents. Secondly, a Notification Process provides asynchronous events to interested applications when metrics tracked by NWS are expected to exceed established normal ranges. Finally, an extended SNMP agent daemon has been developed to provide SNMP-enabled applications with the ability to retrieve information from NWS without the need for custom programming. The new modules help to position NWS as an integral component for performing QoS monitoring and fault prediction in an SNMP-based fault management architecture. We also demonstrate that the additional functionality adds a minimal amount of overhead to the system, and reported bandwidth utilization correlates closely with results from the invasive NWS bandwidth sensor.

关键词： parallel processing systems

来源：评论

学校读者我要写书评

暂无评论

Concepts for execution time prediction of 3D GPU rendering

Concepts for execution time prediction of 3D GPU rendering

引用

international symposium on Industrial Embedded systems

作者： Stephan Schnitzer Simon Gansel Frank Dürr Kurt Rothermel Institute of Parallel and Distributed Systems University of Stuttgart Germany System Architecture and Platforms Department Daimler AG Germany

The relevance of graphical functions in vehicular applications has increased significantly during the last years. Modern cars are equipped with multiple displays used by different applications such as speedometer, navigation system, or media players. The recent trend towards hardware consolidation to reduce hardware cost, installation space, and energy consumption, causes graphical 3D applications of different safety-criticality to share a single GPU. This requires effective real-time GPU scheduling concepts to ensure safety and isolation for 3D rendering. Since current GPUs are not preemptible, a deadline-based scheduler must know the GPU execution time of GPU commands in advance. In this work, we present a novel framework to measure and predict the execution time of GPU commands using OpenGL ES 2.0. We present prediction models for the main GPU commands relevant for 3D rendering, namely, FLUSH, CLEAR, and DRAW. For the DRAW command we propose to use the 3D bounding box of the rendered model and apply the vertex shader projection to heuristically estimate the number of fragments rendered. We finally present the implementation and evaluation of our framework, which demonstrates its feasibility and shows that good prediction accuracy can be achieved. In our evaluation using realistic scenarios the absolute prediction error of the DRAW command did not exceed 260 μs.

关键词： Graphics processing units Context Rendering (computer graphics) Predictive models Three-dimensional displays Real-time systems Time measurement

来源：评论

学校读者我要写书评

暂无评论

Parametrically described regular semigroup interconnection networks for large-scale multicomputers

Parametrically described regular semigroup interconnection n...

引用

international symposium on parallel Architectures, Algorithms and Networks (ISPAN)

作者： O.G. Monakhov E.A. Monakhova Institute of Computational Mathematics and Mathematical Geophysics SD RAS Novosibirsk Russia

ISBN: (纸本)0769509363

We propose a new topology for multicomputer networks: parametrically described, regular, and based on semigroups (PRS) networks (or R/sub s/(N, v, g) graphs with the order N, the degree v, the girth g, and the number of equivalence classes s). Many classes of networks such as hypercubes, circulants, cube-connected cycles, etc. are shown to be special cases of the proposed network. We explore the basic structure, topological properties, optimization of parameters and synthesis of optimal networks having the minimal diameter for the given parameters of the graph. Correspondingly, we examine the optimal characteristics with respect to transit delays and structural survival in such networks. The PRS networks reaching the lower bounds on the diameter were synthesized. In some cases, we found that the new network has a better diameter than classes of networks described in the literature provided they have the same vertex and edge complexity.

关键词： Multiprocessor interconnection networks Large-scale systems Computer networks Network topology Hypercubes distributed computing Geophysics computing Network synthesis Computer architecture Mathematics

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：