Structured parallel programming patterns for efficient computation
Programming is now parallel programming. Much as structured programming revolutionized traditional serial programming decades ago, a new kind of structured programming, based on patterns, is relevant to parallel programming today. Parallel computing experts and industry insiders Michael McCool, Arch...
| Otros Autores: | , , |
|---|---|
| Formato: | Libro electrónico |
| Idioma: | Inglés |
| Publicado: |
Amsterdam ; Boston, Mass. :
Elsevier/Morgan Kaufmann
[2012].
|
| Edición: | 1st edition |
| Materias: | |
| Ver en Biblioteca Universitat Ramon Llull: | https://discovery.url.edu/permalink/34CSUC_URL/1im36ta/alma991009628491106719 |
Tabla de Contenidos:
- Front Cover; Structured Parallel Programming: Patterns for Efficient Computation; Copyright; Table of Contents; Listings; Preface; Preliminaries; 1 Introduction; 1.1 Think Parallel; 1.2 Performance; 1.3 Motivation: Pervasive Parallelism; 1.3.1 Hardware Trends Encouraging Parallelism; 1.3.2 Observed Historical Trends in Parallelism; 1.3.3 Need for Explicit Parallel Programming; 1.4 Structured Pattern-Based Programming; 1.5 Parallel Programming Models; 1.5.1 Desired Properties; 1.5.2 Abstractions Instead of Mechanisms; 1.5.3 Expression of Regular Data Parallelism; 1.5.4 Composability
- 1.5.5 Portability of Functionality1.5.6 Performance Portability; 1.5.7 Safety, Determinism, and Maintainability; 1.5.8 Overview of Programming Models Used; Cilk Plus; Threading Building Blocks (TBB); OpenMP; Array Building Blocks (ArBB); OpenCL; 1.5.9 When to Use Which Model?; 1.6 Organization of this Book; 1.7 Summary; 2 Background; 2.1 Vocabulary and Notation; 2.2 Strategies; 2.3 Mechanisms; 2.4 Machine Models; 2.4.1 Machine Model; Instruction Parallelism; Memory Hierarchy; Virtual Memory; Multiprocessor Systems; Attached Devices; 2.4.2 Key Features for Performance; Data Locality
- Parallel Slack2.4.3 Flynn's Characterization; 2.4.4 Evolution; 2.5 Performance Theory; 2.5.1 Latency and Throughput; 2.5.2 Speedup, Efficiency, and Scalability; 2.5.3 Power; 2.5.4 Amdahl's Law; 2.5.5 Gustafson-Barsis' Law; 2.5.6 Work-Span Model; 2.5.7 Asymptotic Complexity; 2.5.8 Asymptotic Speedup and Efficiency; 2.5.9 Little's Formula; 2.6 Pitfalls; 2.6.1 Race Conditions; 2.6.2 Mutual Exclusion and Locks; 2.6.3 Deadlock; 2.6.4 Strangled Scaling; 2.6.5 Lack of Locality; 2.6.6 Load Imbalance; 2.6.7 Overhead; 2.7 Summary; I Patterns; 3 Patterns; 3.1 Nesting Pattern
- 3.2 Structured Serial Control Flow Patterns3.2.1 Sequence; 3.2.2 Selection; 3.2.3 Iteration; 3.2.4 Recursion; 3.3 Parallel Control Patterns; 3.3.1 Fork-Join; 3.3.2 Map; 3.3.3 Stencil; 3.3.4 Reduction; 3.3.5 Scan; 3.3.6 Recurrence; 3.4 Serial Data Management Patterns; 3.4.1 Random Read and Write; 3.4.2 Stack Allocation; 3.4.3 Heap Allocation; 3.4.4 Closures; 3.4.5 Objects; 3.5 Parallel Data Management Patterns; 3.5.1 Pack; 3.5.2 Pipeline; 3.5.3 Geometric Decomposition; 3.5.4 Gather; 3.5.5 Scatter; 3.6 Other Parallel Patterns; 3.6.1 Superscalar Sequences; 3.6.2 Futures
- 3.6.3 Speculative Selection3.6.4 Workpile; 3.6.5 Search; 3.6.6 Segmentation; 3.6.7 Expand; 3.6.8 Category Reduction; 3.6.9 Term Graph Rewriting; 3.7 Non-Deterministic Patterns; 3.7.1 Branch and Bound; 3.7.2 Transactions; 3.8 Programming Model Support for Patterns; 3.8.1 Cilk Plus; Nesting, Recursion, Fork-Join; Reduction; Map, Workpile; Scatter, Gather; 3.8.2 Threading Building Blocks; Nesting, Recursion, Fork-Join; Map; Workpile; Reduction; Scan; Pipeline; Speculative Selection, Branch and Bound; 3.8.3 OpenMP; Map, Workpile; Reduction; Fork-Join
- Stencil, Geometric Decomposition, Gather, Scatter
