Introduction to Parallel Processing: Algorithms and Architectures下载-CSDN社区

THE CONTEXT OF PARALLEL PROCESSING The field of digital computer architecture has grown explosively in the past two decades. Through a steady stream of experimental research, tool-building efforts, and theoretical studies, the design of an instruction-set architecture, once considered an art, has been transformed into one of the most quantitative branches of computer technology. At the same time, better understanding of various forms of concurrency, from standard pipelining to massive parallelism, and invention of architectural structures to support a reasonably efficient and user-friendly programming model for such systems, has allowed hardware performance to continue its exponential growth. This trend is expected to continue in the near future. This explosive growth, linked with the expectation that performance will continue its exponential rise with each new generation of hardware and that (in stark contrast to software) computer hardware will function correctly as soon as it comes off the assembly line, has its down side. It has led to unprecedented hardware complexity and almost intolerable dev- opment costs. The challenge facing current and future computer designers is to institute simplicity where we now have complexity; to use fundamental theories being developed in this area to gain performance and ease-of-use benefits from simpler circuits; to understand the interplay between technological capabilities and limitations, on the one hand, and design decisions based on user and application requirements on the other.

Behrooz_Parhami - Introduction to Parallel Processing Algorithms and Architectures.pdf

Increasingly, parallel processing is being seen as the only cost-effective method for the fast solution of computationally large and data-intensive problems. The emergence of inexpensive parallel computers such as commodity desktop multiprocessors and clusters of workstations or PCs has made such parallel methods generally applicable, as have software standards for portable parallel programming. This sets the stage for substantial growth in parallel software. Data-intensive applications such as transaction processing and information retrieval, data mining and analysis and multimedia services have provided a new challenge for the modern generation of parallel platforms. Emerging areas such as computational biology and nanotechnology have implications for algorithms and systems development, while changes in architectures, programming models and applications have implications for how parallel platforms are made available to users in the form of grid-based services. This book takes into account these new developments as well as covering the more traditional problems addressed by parallel computers.Where possible it employs an architecture-independent view of the underlying platforms and designs algorithms for an abstract model. Message Passing Interface (MPI), POSIX threads and OpenMP have been selected as programming models and the evolving application mix of parallel computing is reflected in various examples throughout the book.

Title: Handbook of Big Data Technologies Length: 895 pages Edition: 1st ed. 2017 Language: English Publisher: Springer Publication Date: 2017-03-26 ISBN-10: 3319493396 ISBN-13: 9783319493398 Table of Contents Part I Fundamentals of Big Data Processing Big Data Storage and Data Models 1 Storage Models 2 Data Models Big Data Programming Models 1 MapReduce 2 Functional Programming 3 SQL-Like 4 Actor Model 5 Statistical and Analytical 6 Dataflow-Based 7 Bulk Synchronous Parallel 8 High Level DSL 9 Discussion and Conclusion Programming Platforms for Big Data Analysis 1 Introduction 2 Requirements of Big Data Programming Support 3 Classification of Programming Platforms 4 Major Existing Programming Platforms 5 A Unifying Framework 6 Conclusion and Future Directions Big Data Analysis on Clouds 1 Introduction 2 Introducing Cloud Computing 3 Cloud Solutions for Big Data 4 Systems for Big Data Analytics in the Cloud 5 Research Trends 6 Conclusions Data Organization and Curation in Big Data 1 Big Data Indexing Techniques 2 Data Organization and Layout Techniques 3 Non-traditional Workloads in Big Data 4 Curation and Metadata Management in Big Data 5 Conclusion Big Data Query Engines 1 Introduction 2 Massively Parallel Query Engines 3 Hadoop Query Engines 4 SQL on Hadoop 5 Query Optimization 6 Query Execution 7 Summary Large-Scale Data Stream Processing Systems 1 Introduction 2 Programming Models 3 System Support for Distributed Data Streaming 4 Case Study: Stream Processing with Apache Flink 5 Applications, Trends and Open Challenges 6 Conclusions and Outlook Part II Semantic Big Data Management Semantic Data Integration 1 An Important Challenge 2 Current State-of-the-Art 3 The Path Forward Linked Data Management 1 Introduction 2 Background Information 3 Native Linked Data Stores 4 Provenance for Linked Data Non-native RDF Storage Engines 1 Introduction 2 Storing Linked Data Using Relational Databases 3 No-SQL Stores 4 Massively Parallel Processing for Linked Data Exploratory Ad-Hoc Analytics for Big Data 1 Exploratory Analytics for Big Data 2 A Top-K Entity Augmentation System 3 DrillBeyond -- Processing Open World SQL 4 Summary and Future Work Pattern Matching Over Linked Data Streams 1 Overview 2 Linked Data Dissemination System 3 Experimental Evaluation 4 Related Work 5 Summary Searching the Big Data: Practices and Experiences in Efficiently Querying Knowledge Bases 1 Introduction 2 Background 3 The Framework of Cache-Based Knowledge Base Querying 4 Similar Queries Suggestion 5 Cache Replacement 6 Implementation and Experimental Evaluation 7 Related Work 8 Discussion and Conclusion Part III Big Graph Analytics Management and Analysis of Big Graph Data: Current Systems and Open Challenges 1 Introduction 2 Graph Databases 3 Graph Processing 4 Graph Dataflow Systems 5 Gradoop 6 Comparison 7 Current Research and Open Challenges 8 Conclusions and Outlook Similarity Search in Large-Scale Graph Databases 1 Introduction 2 Preliminaries 3 The Pruning-Verification Framework 4 State-of-the-Art Approaches 5 Future Research Directions 6 Summary Big-Graphs: Querying, Mining, and Beyond 1 Introduction 2 Graph Data Models 3 Pattern Matching Techniques Over Big-Graphs 4 Mining Techniques Over Big-Graphs 5 Open Problems 6 Conclusions 7 About Authors Link and Graph Mining in the Big Data Era 1 Introduction 2 Definitions 3 Temporal Evolution 4 Link Prediction 5 Community Detection 6 Graphs in Big Data 7 Weighted Networks 8 Extending Graph Models: Multilayer Networks 9 Open Challenges 10 Conclusions Granular Social Network: Model and Applications 1 Introduction 2 Preliminaries 3 Literature Review 4 Fuzzy Granular Social Networks (FGSN) 5 Discussions and Conclusions Part IV Big Data Applications Big Data, IoT and Semantics 1 Introduction 2 Semantics for Big Data 3 Big Data and Semantics in the Internet of Things 4 Social Mining 5 Graph Mining 6 Big Stream Data Mining 7 Geo-Referenced Data Mining 8 Conclusion SCADA Systems in the Cloud 1 Introduction 2 Related Work 3 An Overview of SCADA 4 Moving SCADA to the Cloud 5 Conceptual SCADA Cloud Orchestration Framework 6 Results 7 Conclusion Quantitative Data Analysis in Finance 1 Introduction 2 The Three V's of Big Data in High Frequency Data 3 Data Cleaning, Aggregating and Management 4 Modeling High Frequency Data in Finance 5 Portfolio Selection and Evaluation 6 The Future 7 Conclusion Emerging Cost Effective Big Data Architectures 1 Introduction 2 Emerging Solutions for Big Data 3 Future Directions 4 Conclusion Bringing High Performance Computing to Big Data Algorithms 1 Introduction 2 GPU Acceleration of Alternating Least Squares 3 GPU Acceleration of Singular Value Decomposition 4 Conclusions Cognitive Computing: Where Big Data Is Driving Us 1 Cognitive Computing: An Alternative Approach for Clear Understanding 2 Big Data Impulsing Cognitive System 3 Traditional Systems versus Cognitive Systems? 4 Data Mining in the Era of Cognitive Systems 5 Design Methods for Cognitive Systems 6 Cognitive Systems 7 The Future of Cognitive Systems 8 Final Remarks Privacy-Preserving Record Linkage for Big Data: Current Approaches and Research Challenges 1 Introduction 2 Background 3 Privacy Aspects and Techniques for PPRL 4 Scalability Techniques for PPRL 5 Multi-party PPRL 6 Open Challenges 7 Conclusions

Preface About the Editors Part 1—Fundamentals and Neuro-Fuzzy Signal Processing Chapter 1—Fuzzy Logic and Neuro-Fuzzy Systems in Medicine and Bio-Medical Engineering: A Historical Perspective 1. The First Period: The Infancy 2. Further Developments and Background 3. Neuro-Fuzzy Systems and their Applications in Medicine and Biology 4. Genetic Algorithms, Fuzzy Logic, and Neuro-Fuzzy Systems 5. Bibliographies 6. Conclusions and Predictions Chapter 2—The Brain as a Fuzzy Machine: A Modeling Problem 1. The Fuzzy Approach in Neurobiology: A Historical Perspective 2. The Generality of Young’s Hypothesis 2.1 Simple Stimuli 2.2 Neural Organization of Cryptic Events: From Tastes to Faces 2.2.1 The General Approach 2.2.2 Solutions Possible: Taste 2.2.3 Solutions Possible: Faces 2.3 Neural Codes 3. Fuzzy Models for Taste 3.1 Grades of Membership in Fuzzy Sets 3.2 A Fuzzy Model 3.2.1 The Model 3.2.2 The Synthesis of the Fuzzy Model 3.2.3 Simulating the Dynamics of Taste Neurons 4. Fuzzy Model For Brain Activity 4.1 A Neural Network Implementing a Fuzzy Machine? 4.2 An Artificial Neuron Implements a Fuzzy Membership Function 4.3 A Layer of Neurons Implements a Fuzzifier 4.4 A “Hidden” Neuron Implements a Fuzzy Rule 5. Applications of Fuzzy Logic to Neural Systems 5.1 Quantitative Aspects of the Fuzzy Neural Sets 5.1.1 Neural Mass 5.1.2 Sensitivity to Fine Gradations in Input 5.1.3 Intelligence 5.2 Defuzzification and Responses 5.3 Memory: Input and Retrieval 6. Conclusions Appendix 1. Abbreviations Appendix 2. Terminology References Chapter 3—Brain State Identification and Forecasting of Acute Pathology Using Unsupervised Fuzzy Clustering of EEG Temporal Patterns 1. Introduction 2. Background 2.1 The Electroencephalogram (EEG) Signal [1], [2] 2.2 Brain States and the EEG 2.3 Stimulus-Evoked EEG Patterns 2.4 Underlying Processes 2.5 Fuzzy Systems and the EEG 3. Tools 3.1 Data Acquisition 3.1.1 Spontaneous Ongoing Signal 3.1.2 Evoked Responses 3.2 Feature Extraction 3.2.1 Spectrum Estimation 3.2.2 Time-Frequency Analysis 3.2.2.1 Multiscale Decomposition By The Fast Wavelet Transform 3.2.2.2 Multichannel Model-Based Decomposition by Matching Pursuit 3.3 The Unsupervised Optimal Fuzzy Clustering (UOFC) Algorithm. 3.4 The Weighted Fuzzy K-Mean (WFKM) Algorithm 3.5 The Clustering Validity Criteria 4. Examples of Uses 4.1 Sleep-Stage Scoring 4.2 Forecasting Epilepsy 4.3 Classifying Evoked and Event-Related Potentials by Waveform 5. Concluding Remarks and Future Applications 5.1 Dynamic Version of State Identification by UOFC 5.2 Data Fusion Appendiex 1: The Fast Wavelet Transform Appendix 2: Multichannel Model-Based Decomposition by Matching Pursuit Appendix 3: Feature Extraction and Reduction by Principal Component Analysis List of Acronyms References Chapter 4—Contouring Blood Pool Myocardial Gated SPECT Images with a Sequence of Three Techniques Based on Wavelets, Neural Networks, and Fuzzy Logic 1. Introduction 2. Anatomy of the G-SPECT Images 3. Strategy of the Proposed Method 3.1. Overview of the Method 3.2. Wavelets-Based Image Pre-Processing 3.3. Neural Network Based Image Segmentation 3.4. Fuzzy Logic-Based Recognition of the Regions of Interest (Ventricles) 3.4.1. Definition of the Required Fuzzy Sentences 3.4.2. Combining Neuronal Approaches and Fuzzy Logic-Based Inference Systems 3.5. Training the Recognition System Using a Neuro-Fuzzy Technique 3.5.1. Automated Generation of Rules and Membership Functions (ALGORAM) 3.5.2. Adjustment of Membership Functions Using a Descent Method (FUNNY) 3.5.3. Combining the Automated Generation of Rules and Membership Functions and the Adjustment of their Parameters in a Parallel Implementation (FUNNY-ALGORAM) 4. In Vitro Experiments and Application to Medical Cases 4.1. Experiments with Phantoms 4.2. Clinical Test Cases 4.3. Implementation Issues 5. Conclusions References Chapter 5—Unsupervised Brain Tumor Segmentation Using Knowledge-Based Fuzzy Techniques 1. Introduction 2. Domain Background 2.1 Slices of Interest for the Study 2.2 Basic MR Contrast Principles 2.3 Knowledge-Based Systems 2.4 System Overview 3. Classification Stages 3.1 Stage Zero: Pathology Detection 3.2 Stage One: Building the Intra-Cranial Mask 3.3 Stage Two: Multi-spectral Histogram Thresholding 3.4 Stage Three: “Density Screening” in Feature Space 3.5 Stage Four: Region Analysis and Labeling 3.5.1 Removing Meningial Regions 3.5.2 Removing Non-Tumor Regions 3.6 Stage Five: Final T1 Threshold 4. Results 4.1 Knowledge-Based vs. Supervised Methods 4.2 Evaluation Over Repeat Scans 5. Discussion References Abbreviations Part 2—Neuro-Fuzzy Knowledge Processing Chapter 6—An Identification of Handling Uncertainties Within Medical Screening: A Case Study Within Screening for Breast Cancer 1. Introduction 2. Screening 2.1 Notations 2.2 The Screening Program 2.3 The Methods 3. The Select Function 3.1 The Decision Step 3.2 Disease-Specific Knowledge 3.3 The Refinement Step 4. A Breast Cancer Case Study 4.1 Minimizing A0 as Much as Possible in One Step 4.2 Finding the Screening Method 4.3 Defining Disease-Specific Knowledge 4.4 Performing the Refinement 4.5 The Integrated System 5. Conclusions and Further Work References Chapter 7—A Fuzzy System For Dental Developmental Age Evaluation 1. Introduction 2. Technical Consideration 2.1 Basic Conception of the Teeth Evaluation System 2.2 Rule Evaluation Module 3. System Optimization by Using Clinical Data 3.1 Material and Method 3.2 Dimensionality Analysis by Principal Component Analysis 3.3 System Optimization by Using Genetic Algorithm 3.4 System Evaluation and Results 4. Discussion and Conclusions Chapter 8—Fuzzy Expert System For Myocardial Ischemia Diagnosis 1. Introduction 2. Fuzzy Expert Systems 3. Difus - Hierarchical Diagnosis Fuzzy System 3.1 Characteristics 3.2 Knowledge Organization 3.3 Structure 3.4 Operation 4. Multimethod Myocardial Ischemia Diagnosis 5. Multimethod Myocardial Ischemia Diagnosis System 5.1 The Implementation of Fuzzy Score-Based Tests 5.1.1 Medical Patterns 5.1.2 Sequential Processing 5.1.3 Compact Representation of Fuzzy Score-Based Tests 5.2 MMIDS Structure and Operation 5.2.1 MMIDS Secondary Group 5.2.2 MMIDS Primary Groups 5.3 Experimental Results 6. Conclusions References Chapter 9—Design and Tuning of Fuzzy Rule-Based Systems for Medical Diagnosis 1. Introduction 2. Problem Statement and General Methodology 3. Design and Rough Tuning of Fuzzy Rules 3.1. Matrix of Knowledge 3.2. Fuzzy Model with Discrete Output 3.3. Fuzzy Model with Continuous Output 3.4. Rough Tuning of Fuzzy Rules 3.4.1. Rough Tuning of Membership Functions 3.4.2. Rough Tuning of Rules Weights 4. Fine Tuning of the Fuzzy Rules with Continuous Output 4.1. Tuning as a Problem of Optimization 4.2. Quality Evaluation of Fuzzy Inference 4.3. Computer Simulation 4.3.1. Experiment 1 4.3.2. Experiment 2 5. Fine Tuning of the Fuzzy Rules with Discrete Output 5.1. Tuning as a Problem of Optimization 5.2. Quality Evaluation of Fuzzy Inference 5.3. Computer Simulation 6. Application to Differential Diagnosis of Ischemia Heart Disease 6.1. Diagnosis Types and Parameters of Patient’s State 6.2. Fuzzy Rules 6.3. Fuzzy Logic Equation 6.4. Rough Membership Functions 6.5. Algorithm of Decision Making 6.6. Fine Tuning of The Fuzzy Rules in Medical Applications 7. Conclusions References Appendix 1—Comparison of real and inferred decisions for 65 patients Appendix 2—FUZZY EXPERT Shell and its Application Chapter 10—Integration of Medical Knowledge in an Expert System for Use in Intensive Care Medicine 1. Introduction 2. Software Design Principles 3. Medical Knowledge in Intensive Care Medicine 3.1. Structure of the Knowledge 3.2. Meaning of Colloquial Rules 3.3. Rule Processing and Result Calculation 3.4. Combining Different Rules 4. Transformation of Knowledge into FLORIDA Commands 4.1. Introduction 4.2. Comments 4.3. Modules 4.4. Linguistic Variables 4.5. The FLORIDA Calculator 4.6. Rules: The Knowledge Itself 4.7. Changing the Normal Value 5. Invocation of FLORIDA 6. Explaining More of FLORIDA’s Functionality — The Knowledge Base Inflammation 6.1. Structuring the Knowledge 6.2. Rules for Fever 6.3. Rules for Leukocytosis/Leukopenia 6.4. Rules for Tachycardia/Tachypnoe 6.5. Rules for Synthesis of Acute Phase Proteins 6.6. Rules for Consumption of Coagulation Components 6.7. Improvement of Explanation 7. Differentiation of Dysfunctions 8. Visualization of the Result 9. Discussion and Conclusions References Part 3—Neuro-Fuzzy Control and Hardware Chapter 11—Hemodynamic Management with Multiple Drugs using Fuzzy Logic 1. Introduction 1.1 Progress in Decision Making 1.2 Progress in Control 2. System Development 2.1 Decision-Making: Fuzzy Decision-Making Module (FDMM) 2.1.1 Purpose 2.1.2 Operation 2.2 Drug-Titration Control: Fuzzy Hemodynamic Control Module (FHCM) 2.2.1 Purpose 2.2.2 Operation 2.3 Supervisory Commands: Therapeutic Assessment Module (TAM) 2.4 System Evaluations 2.4.1 Example One 2.4.2 Example Two 3. Future Prospects 3.1 Design Possibilities 3.2 “Curse of Dimensions” 3.3 Machine Intelligence Additional Resources Appendix: Terminology References Chapter 12—Neuro-Fuzzy Hardware in Medical Applications 12 A.—System Requirements for Fuzzy and Neuro-Fuzzy Hardware in Medical Equipment 1. Introduction 2. Specific Requirements of Medical Applications 2.1 General System and Technological Requirements 2.2 Reliability Requirements 2.3 Precision and Sensitivity to Parameters 3. Analysis of Several Applications 3.1 Life-Support Applications 3.1.1 Artificial Heart Control 3.1.2 Assisted Ventilation 3.2 Anesthesia Related Equipment 3.3 Fuzzy and Neuro-Fuzzy-Based Equipment for Prosthetics 3.4 General Purpose Devices 3.5 Other Applications 4. General System Design Issues 4.1 Nonlinearity Implementation - Simulation Power 4.2 Dynamic Errors 5. Hardware Implementation Issues 5.1 Implementation Choice: Analog vs. Digital Fuzzy Processors 5.2 Hardware Minimization 5.3 Parallelism vs. Number of Rule Blocks 5.4 A Minimal System Design 6. Choosing the Right Design 7. Conclusions References Chapter 12 B—Neural Networks and Fuzzy-Based Integrated Circuit and System Solutions Applied to the Biomedical Field 1. Introduction 2. Required Properties for Embedded Medical Systems 2.1 Embedding medical systems 2.2 Autonomy 2.3 Reliability - safety 2.4 Precision of computation 2.5 Application-specific requirements 3. Architectures Applied to Neuro-Fuzzy IC Design 3.1 Artificial Neural Network Integrated Realization 3.2 Fuzzy-Based Integrated Realization 3.3 Hybrid Integrated Realization 3.4 An example of neuro-fuzzy realization 4. Concluding Remarks References Index of Terms