检索结果-内蒙古大学图书馆

A simple hardware prefetching scheme using sequentiality for shared-memory multiprocessors

IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS 1997年第11期E80D卷 1055-1063页

作者： Tcheun, MK Maeng, SR Cho, JW Dept. of Computer Science and CAIR Korea Adv. Inst. Sci. and Technol. Yusung-Gu Taejon 305-701 373-1 Kusung-Dong South Korea Korea Adv. Inst. Sci. and Technol. South Korea Dept. Comp. Sci. Korea Adv. Inst. S. South Korea Dept. Comp. Sci. Korea Adv. Inst. S. Ctr. for Artif. Intell. Research Member Asian Activities Comm. I. South Korea

To reduce the memory access latency on shared-memory multiprocessors, several prefetching schemes have been proposed. The sequential prefetching scheme is a simple hardware-controlled scheme, which exploits the sequentiality of memory accesses to predict which blocks will be read in the near Future. Aggressive sequential prefetching prefetches many blocks on each miss to reduce the miss rates and results in good performance For application programs with high sequentiality. However, conservative sequential prefetching prefetches a few blocks on each miss to avoid prefetching of useless blocks, which shows better performance than aggressive sequential prefetching for application programs with low sequentiality. We analyze the relationship between the sequentiality of application programs and the effectiveness of sequential prefetching on various memory and network latency and propose a new adaptive sequential prefetching scheme. Simply adding a small table to the sequential prefetching scheme, the proposed scheme prefetches a large number of blocks for application programs with high sequentiality and reduces the miss rates significantly, and prefetches a small number of blocks for application programs with low sequentiality and avoids loading useless blocks.

关键词： shared-memory multiprocessors sequential prefetching sequentiality prefetching degree

来源：评论

学校读者我要写书评

暂无评论

Partitioning Processor Arrays under Resource Constraints

引用

Journal of VLSI Signal Processing Systems for Signal, Image, and Video Technology 1997年第1期17卷 5-20页

作者： Teich, Jürgen Thiele, Lothar Zhang, Lee Z. Comp. Eng. and Networks Laboratory Swiss Fed. Institute of Technology CH-8092 Zurich Switzerland Institute on Microelectronics University of Saarland D-66041 Saarbrücken Germany University of Kaiserslautern Germany University of Saarland Germany Dept. of Elec. Eng. and Comp. Sci. UC Berkeley United States Inst. Tech. Comp. Sci. Commun. N. ETH Zurich Switzerland IEEE Switzerland Technical University of Munich Germany Inst. Netwk. Theor. and Circuit Des. Technical University Munich Germany Information Systems Laboratory Stanford University United States Faculty of Engineering University of Saarland Saarbrucken Germany ETH Zurich Switzerland Technical University Switzerland IEEE Circuits and Systems Society United States Tsinghua University Beijing China University of the Saarland Saarbrucken Germany Phase 5 Company Oberursel Germany

A single integer linear programming model for optimally scheduling partitioned regular algorithms is presented. The herein presented methodology differs from existing methods in the following capabilities: 1) Not only constraints on the number of available processors and communication capabilities are taken into account, but also local memories and constraints on the size of available memories. 2) Different types of processors can be handled. 3) The size of the optimization model (number of integer variables) is independent of the size of the tiles to be executed. Hence, 4) the number of integer variables in the optimization model is greatly reduced such that problems of relevant size can be solved in practical execution time.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Vorticity-velocity formulation for the Navier-Stokes equations: The three-dimensional case

引用

Numerical Methods for Partial Differential Equations 1997年第1期13卷 19-19页

作者： Medjo, T.T. Inst. for Appl. Math. and Sci. Comp. Indiana University Bloomington IN 47405 United States Lab. d'Analyse Numérique Batiment 425 Université de Paris-Sud 91400 Orsay France

In this article, we propose a mixed method for the vorticity-velocity formulation of the stationary Stokes and Navier-Stokes equations in space dimension three, the unknowns being the vorticity and the velocity of the fluid. We give a similar variational formulation for the nonstationary Stokes equations in the vorticity-velocity variables. © 1997 John Wiley & Sons, Inc.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Measurement of sodium fluorescein wash-in time constants in subjects with peripheral vascular disease

引用

Biomedical inst.umentation and Technology 1997年第6期31卷 600-607页

作者： Oh, Deborah K. Zhang, Anjing Magin, Richard L. Ctr. for Biophys. and Compl. Biology University of Illinois Beckman Inst. Adv. Sci. and Technol. Urbana IL United States Dept. of Elec. and Comp. Engineering University of Illinois Beckman Inst. Adv. Sci. and Technol. Urbana IL United States Gandalf Canada Ltd. Nepean Ont. Canada Beckman Inst. Adv. Sci. and Technol. 405 North Mathews Urbana IL 61801 United States

The authors developed a noninvasive two-channel dynamic dermofluorometer that can quantitatively follow the rapid skin wash-in kinetics of a fluorescent dye to provide an assessment of local skin perfusion. The dermofluorometer was tested in normal subjects and diabetic patients with and without peripheral vascular disease. After an intravenous injection of 1-2 mL of a 10% solution of sodium fluorescein (1.1-2.8 mg/kg), the fluorescent signal was monitored from two sites on the skin surfaces of the forearm and foot. A 3.2-mm-diameter glass fiberoptic bundle was used both to transmit the excitation light (489 nm) and to receive the fluorescent emission (517 nm). Dermofluorometer readings were recorded approximately every second for 10-15 minutes following the injection. The time course of the fluorescein signal intensity was fit to a single exponential curve characterized by a wash-in time constant. There was no significant difference in arm wash-in time constants. Foot wash-in time constants were increased in diabetic patients who had past histories of foot ulcers relative to diabetic patients without a history of foot ulcers (3.2 vs 1.6 min., p < 0.05). Foot wash-in time constants were decreased in diabetic patients who had active infected foot ulcers. This study demonstrates the ability of the dynamic dermofluorometer to measure wash-in constants that reflect the local skin perfusion in less than 15 minutes after a low intravenous dose of sodium fluorescein.

关键词： Fluorometers

来源：评论

学校读者我要写书评

暂无评论

On distributed representation of output layer for recognizing Japanese Kana characters using neural networks

On distributed representation of output layer for recognizin...

引用

International Conference on Document Analysis and Recognition

作者： K. Kenni H. Shimodaira K. Nakayama Sch. of Inf. Sci. Japan Adv. Inst. of Sci. & Technol. Ishikawa Japan School of Information Science Japan Advanced Institute of Science and Technology Tatsunokuchi Ishikawa Japan Department of Elec. & Comp. Eng. Kanazawa University Kanazawa Japan

The paper presents an automatic coding scheme for representing the output layer of a neural network. comp.red to local representation where the number of output unit is p, the number of output unit required for the proposed representation is close to logp. The output of seven different printers were used for evaluating the performance of the system. The proposed automatic representation gave the average recognition rate of 98.7% for 71 categories.

关键词： Character recognition Neural networks Printers Image databases Information sci.nce Feature extraction Data preprocessing Testing National electric code

来源：评论

学校读者我要写书评

暂无评论

Information technology in three small developed countries

引用

Journal of Management Information Systems 1997年第4期13卷 61-89页

作者： Ein-Dor, P. Myers, M.D. Raman, K.S. Department of Information Systems Faculty of Management Tel Aviv University Israel Marcel Annie Adams Inst. Bus. M. Dept. of Mgmt. Sci. and Info. Syst. University of Auckland New Zealand Dept. of Info. Syst. and Comp. Sci. National University of Singapore Info. Systems Area of the Department

The perception that small size may no longer be an economic disadvantage to either organizations or countries has become fairly widespread. We share the perception that not only are small countries not at a disadvantage, but they may actually have an advantage over larger comp.titors with regard to information technology industries. Indeed, small developed countries are achieving considerable success in the development of their information technology industries. The three countries studied here-Israel, New Zealand, and Singapore-are among those that have seen the rapid development of their IT industries in recent years, despite the fact that they are among the world's smaller countries, are considerably different in many geographic, cultural, and political respects, and are geographically dispersed around the globe. The objective of this study is to describe and comp.re the information technology industries of these three small developed countries. All three countries, with no inherent advantages in raw materials and only small domestic markets, have the IT infrastructure and the human skills needed. They are nimble and flexible and can find niche markets in which to specialize. Not all three, however, have developed IT industries to the same degree. There could be factors specific to small developed countries that facilitate the development of indigenous IT production. The dominant factor that seems to provide some explanation for different levels and directions of development of IT production is government policy in promoting IT production directly, in supporting IT industry R and D, and in education policies designed to provide appropriately trained labor pools.

关键词： Factors of national comp.ting Information systems in developed countries International comp.ting National comp.ting policies

来源：评论

学校读者我要写书评

暂无评论

Automated Discovery of Active Motifs in Three Dimensional Molecules 3

Automated Discovery of Active Motifs in Three Dimensional Mo...

引用

3rd International Conference on Knowledge Discovery and Data Mining, KDD 1997

作者： Wang, Xiong Wang, Jason T.L. Shasha, Dennis Shapiro, Bruce Dikshitulu, Sitaram Rigoutsos, Isidore Zhang, Kaizhong Comp. & Infor. Sci. New Jersey Inst. Tech. NewarkNJ07102 United States Courant Institute New York University New YorkNY10012 United States Lab. of Math. Biology National Cancer Institute FrederickMD21702 United States Data/Knowledge Eng. Lab. New Jersey Inst. Tech. NewarkNJ07102 United States IBM Watson Research Ctr. Yorktown HeightsNY10598 United States Univ. of Western Ontario OntarioN6A 5B7 Canada

ISBN: (纸本)1577350278

In this paper we present a method for discovering approximately common motifs (also known as active motifs) in three dimensional (3D) molecules. Each node in a molecule is represented by a 3D point in the Euclidean Space and each edge is represented by an undirected line segment connecting two nodes in the molecule. Motifs are rigid substructures which may occur in a molecule after allowing for an arbitrary number of rotations and translations as well as a small number (specified by the user) of node insert/delete operations in the motifs or the molecule. (We call this "approximate occurrence.") The proposed method combines the geometric hashing technique and block detection algorithms for undirected graphs. To demonstrate the utility of our algorithms, we discuss their applications to classifying three families of molecules pertaining to antibacterial sulfa drugs, anti-anxiety agents (benzodiazepines) and antiadrenergic agents (β receptors). Experimental results indicate the good performance of our algorithms and the high quality of the discovered motifs. Copyright © 1997, American Association for Artificial Intelligence (***). All rights reserved.

关键词： Molecules

来源：评论

学校读者我要写书评

暂无评论

algorithms for Hierarchical Spatial Reasoning

引用

GeoInformatica 1997年第3期1卷 251-273页

作者： Papadias, Dimitris Egenhofer, Max J. Department of Computer Science Hong Kong Univ. of Sci. and Technol. Clearwater Bay Hong Kong Natl. Ctr. Geogr. Info. and Anal. Department of Computer Science University of Maine Orono ME 04469-5711 United States Natl. Ctr. Geogr. Info. and Anal. University of Maine Germany Dept. of Spatial Info. Sci. and Eng. United States University of Stuttgart Germany Univ. of Maine in Surv. Engineering United States Computer Science Dept. Hong Kong Univ. of Sci. and Technol. Hong Kong Inst. Intelligent Robotics Info. S. Canada Elec. and Computer Engineering Dept. Natl. Technical University of Athens Greece Department of Geoinformation Technical University of Vienna Austria Dept. of Comp. Sci. and Engineering University of California San Diego CA United States Natl. Ctr. Geogr. Info. and Anal. University of Maine United States Artif. Intell. Research Division Ger. Natl. Res. Ctr. Info. Technol. Germany

In several applications, there is the need to reason about spatial relations using multiple local frames of reference that are hierarchically organized. This paper focuses on hierarchical reasoning about direction relations, a special class of spatial relations that describe order in space (e.g., north or northeast). We assume a spatial database of points and regions. Points belong to regions, which may recursively be parts of larger regions. The direction relations between points in the same region are explicitly represented (and not calculated from coordinates). Inference mechanisms are applied to extract direction relations between points located in different regions and to detect inconsistencies. We study two comp.ementary types of inference. The first one derives the direction relation between points from the relations of their ancestor regions. The second type derives the relation through chains of common points using path consistency. We present algorithms for both types of inference and discuss their comp.tational comp.exity.

关键词： Direction relations Geographic databases Path consistency Spatial reasoning

来源：评论

学校读者我要写书评

暂无评论

Visibility comp.tations on Hierarchical Triangulated Terrain Models

引用

GeoInformatica 1997年第3期1卷 219-250页

作者： De Floriani, Leila Magillo, Paola Dipto. Info. Sci. dell'Informazione Università di Genova 16146 Genova Via Dodecanese 35 Italy Department of Computer Science University of Genova Italy University of Genova Italy Institute for Applied Mathematics Italian National Research Council Genova Italy Department of Mathematics University of Genova Italy Institute of Applied Mathematics Italian National Research Council Italy ACM IEEE Computer Society Intl. Assoc. for Pattern Recognition Italy University of Genova Italy Inst. Natl. Rech. Info. et Automat. Sophia Antipolis France Dept. of Comp. and Info. Sciences University of Genova Italy Intl. Assoc. for Pattern Recognition Italy

Hierarchical terrain models provide a multiresolution description of a topographic surface based on a nested partition of the domain. The tree-like structure of these models is an effective support to processing spatial operations. In this paper, we consider visibility comp.tations on hierarchical terrain models based on triangular subdivisions, called Hierarchical Triangulated Irregular Networks (HTINs). We address two basic problems in visibility comp.tation, namely determining the visibility of a query object, and comp.ting the viewshed of a given viewpoint. We propose algorithms for performing such operations on an HTIN at variable resolution. A general drawback of hierarchical models is in the inconsistency of representations at variable resolution obtained from them, since vertical gaps may occur at edges where different resolutions meet. The algorithms proposed here avoid this undesired effect. A related, but independent, contribution of this paper is also a new algorithm for extracting a consistent terrain representation at variable resolution from an HTIN.

关键词： Data structures and algorithms Hierarchical terrain models Multiresolution representations Visibility comp.tation

来源：评论

学校读者我要写书评

暂无评论

Incremental least squares methods and the extended Kalman filter

引用

SIAM JOURNAL ON OPTIMIZATION 1996年第3期6卷 807-822页

作者： Bertsekas, DP Dept. of Elec. Eng. and Comp. Sci. Massachusetts Inst. of Technology Cambridge MA 02139 United States

In this paper we propose and analyze nonlinear least squares methods which process the data incrementally, one data block at a time. Such methods are well suited for large data sets and real time operation and have received much attention in the context of neural network training problems. We focus on the extended Kalman filter, which may be viewed as an incremental version of the Gauss-Newton method. We provide a nonstochastic analysis of its convergence properties, and we discuss variants aimed at accelerating its convergence.

关键词： optimization least squares Kalman filter

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：