Intel Threading Building Blocks.pdf

发布时间：2022-05-29 发布人：admin 分类：说明书资料大小：2.87M 资料格式：pdf 举报版权申诉

zmm19861210-4859367-4744302542969395379.pdf-第1页.png

第1页 / 共334页

zmm19861210-4859367-4744302542969395379.pdf-第2页.png

第2页 / 共334页

zmm19861210-4859367-4744302542969395379.pdf-第3页.png

第3页 / 共334页

zmm19861210-4859367-4744302542969395379.pdf-第4页.png

第4页 / 共334页

zmm19861210-4859367-4744302542969395379.pdf-第5页.png

第5页 / 共334页

zmm19861210-4859367-4744302542969395379.pdf-第6页.png

第6页 / 共334页

zmm19861210-4859367-4744302542969395379.pdf-第7页.png

第7页 / 共334页

zmm19861210-4859367-4744302542969395379.pdf-第8页.png

第8页 / 共334页

Intel Threading Building Blocks

Table of Contents

Foreword

Note from the Lead Developer of Intel Threading Building Blocks

Preface

Assumptions This Book Makes

Contents of This Book

Conventions Used in This Book

Informal Class Declarations

Using Code Examples

How to Contact Us

Acknowledgments

Why Threading Building Blocks?

Overview

Benefits

Comparison with Raw Threads and MPI

Comparison with OpenMP

Recursive Splitting, Task Stealing, and Algorithms

Thinking Parallel

Elements of Thinking Parallel

Decomposition

Data Parallelism

Task Parallelism

Pipelining (Task and Data Parallelism Together)

Mixed Solutions

Achieving Parallelism

Scaling and Speedup

How Much Parallelism Is There in an Application?

Amdahl’s Law

Gustafson’s observations regarding Amdahl’s Law

What did they really say?

Serial versus parallel algorithms

What Is a Thread?

Programming Threads

Safety in the Presence of Concurrency

Mutual Exclusion and Locks

Correctness

Abstraction

Patterns

Intuition

Basic Algorithms

Initializing and Terminating the Library

Loop Parallelization

parallel_for

Grain size

Automatic grain size

Notes on automatic grain size

parallel_for with partitioner

parallel_reduce

Advanced example

Parallel_reduce with partitioner

Advanced Topic: Other Kinds of Iteration Spaces

Notes on blocked_range2d

parallel_scan

Parallel_scan with partitioner

Recursive Range Specifications

Splittable Concept

Model Types: Splittable Ranges

split Class

Range Concept

Model Types

blocked_range Template Class

blocked_range2d Template Class

Partitioner Concept

Model Types: Partitioners

simple_partitioner Class

auto_partitioner Class

parallel_for Template Function

parallel_reduce Template Function

parallel_scan Template Function

pre_scan_tag and final_scan_tag Classes

Summary of Loops

Advanced Algorithms

Parallel Algorithms for Streams

Cook Until Done: parallel_while

Notes on parallel_while scaling

parallel_while Template Class

Working on the Assembly Line: Pipeline

Throughput of pipeline

Nonlinear pipelines

pipeline Class

filter Class

parallel_sort

parallel_sort Template Function

Containers

concurrent_queue

Iterating over a concurrent_queue for Debugging

When Not to Use Queues

concurrent_queue Template Class

concurrent_vector

concurrent_vector Template Class

Whole Vector Operations

Concurrent Operations

Parallel Iteration

Capacity

Iterators

concurrent_hash_map

Further Reading

Index

Praise for Intel Threading Building Blocks “The Age of Serial Computing is over. With the advent of multi-core processors, parallel- computing technology that was once relegated to universities and research labs is now emerging as mainstream. Intel Threading Building Blocks updates and greatly expands the ‘work-stealing’ technology pioneered by the MIT Cilk system of 15 years ago, providing a modern industrial-strength C++ library for concurrent programming. “Not only does this book offer an excellent introduction to the library, it furnishes novices and experts alike with a clear and accessible discussion of the complexities of concurrency.” — Charles E. Leiserson, MIT Computer Science and Artificial Intelligence Laboratory “We used to say make it right, then make it fast. We can’t do that anymore. TBB lets us design for correctness and speed up front for Maya. This book shows you how to extract the most benefit from using TBB in your code.” — Martin Watt, Senior Software Engineer, Autodesk “TBB promises to change how parallel programming is done in C++. This book will be extremely useful to any C++ programmer. With this book, James achieves two important goals: • Presents an excellent introduction to parallel programming, illustrating the most com- mon parallel programming patterns and the forces governing their use. • Documents the Threading Building Blocks C++ library—a library that provides generic algorithms for these patterns. “TBB incorporates many of the best ideas that researchers in object-oriented parallel computing developed in the last two decades.” — Marc Snir, Head of the Computer Science Department, University of Illinois at Urbana-Champaign “This book was my first introduction to Intel Threading Building Blocks. Thanks to the easy-to-follow discussion of the features implemented and the reasons behind the choices made, the book makes clear that Intel’s Threading Building Blocks are an excellent synthesis of some of the best current parallel programming ideas. The judicious choice of a small but powerful set of patterns and strategies makes the system easy to learn and use. I found the numerous code segments and complete parallel applications presented in the book of great help to understand the main features of the library and illustrate the different ways it can be used in the development of efficient parallel programs.” — David Padua, University of Illinois

“The arrival of the multi-core chip architecture has brought great challenges in parallel programming and there is a tremendous need to have good books that help and guide the users to cope with such challenges. “This book on Intel Threading Building Blocks provides an excellent solution in this direc- tion and is likely to be an important text to teach its readers on parallel programming for multi-cores. “The book illustrates a unique path for readers to follow in using a C++-based parallel programming paradigm—a powerful and practical approach for parallel programming. It is carefully designed and written, and can be used both as a textbook for classroom training, or a cookbook for field engineers.” — Professor Guang R. Gao, University of Delaware “I enjoyed reading this book. It addresses the need for new ways for software developers to create the new generation of parallel programs. In the presence of one of the ‘largest disruptions that information technology has seen’ (referring to the introduction of multi- core architectures), this was desperately needed. “This book also fills an important need for instructional material, educating software engi- neers of the new opportunities and challenges. “The library-based approach, taken by the Threading Building Blocks, could be a signifi- cant new step, as it complements approaches that rely on advanced compiler technology.” — Rudolf Eigenmann, Purdue University, Professor of ECE and Interim Director of Computing Research Institute “Multi-core systems have arrived. Parallel programming models are needed to enable the creation of programs that exploit them. A good deal of education is needed to help sequential programmers adapt to the requirements of this new technology. This book represents progress on both of these fronts. “Threading Building Blocks (TBB) is a flexible, library-based approach to constructing parallel programs that can interoperate with other programming solutions. “This book introduces TBB in a manner that makes it fun and easy to read. Moreover, it is packed full of information that will help beginners as well as experienced parallel programmers to apply TBB to their own programming problems.” — Barbara Chapman, CEO of cOMPunity, Professor of Computer Science at the University of Houston

“Future generations of chips will provide dozens or even hundreds of cores. Writing appli- cations that benefit from the massive computational power offered by these chips is not going to be an easy task for mainstream programmers who are used to sequential algo- rithms rather than parallel ones. “Intel’s TBB is providing a big step forward into this long path, and what is better, all in the C++ framework.” — Eduard Ayguade, Barcelona Supercomputer Center, Technical University of Catalunya “Intel’s TBB is to parallel programming what STL was to plain C++. Generic program- ming with STL dramatically improved C++ programming productivity. TBB offers a generic parallel programming model that hides the complexity of concurrency control. It lowers the barrier to parallel code development, enabling efficient use of ‘killer’ multi- cores.” — Lawrence Rauchwerger, Texas A&M University, Inventor of STAPL “For the last eighteen years the denizens of the thinly populated world of supercomputers have been looking for a way to write really pretty and practical parallel programs in C++. We knew templates and generic programming had to be part of the answer, but it took the arrival of multi-core (and soon many-core) processors to create a fundamental change in the computing landscape. Parallelism is now going to be everyday stuff. “Every C++ programmer is going to need to think about concurrency and parallelism and Threading Building Blocks provides the right abstractions for them to do it correctly. “This book is not just a discussion of a C++ template library. It provides a lovely and in- depth overview of much of what we have learned about parallel computing in the last 25 years. It could be a great textbook for a course on parallel programming.” — Dennis Gannon, Science Director, Pervasive Technology Labs at Indiana University, former head of DARPA’s High Performance Computing (HPC++) project, and steering committee member of the Global Grid Forum “TBB hits the application developer’s sweet spot with such advantages as uniprocessor performance, parallel scalability, C++ programming well beyond OpenMP, compatibility with OpenMP and hand threads, Intel Threading Tools support for performance and confidence, and openness to the software development community. TBB avoids several constraints surrounding the sweet spot: language extension risks, specific compiler depen- dences and hand-threading complexities. “This book should make developers productive without a steep training curve, and the applications they produce should be of high quality and performance.” — David Kuck, Intel Fellow, founder of KAI and former director of the Center for Supercomputing Research and Development

Intel Threading Building Blocks Outfitting C++ for Multi-Core Processor Parallelism

Other resources from O’Reilly Related titles C++ Cookbook™ Practical C++ Programming C++ in a Nutshell Pthreads Programming: A POSIX Standard for Better Multiprocessing Secure Programming Cook- book for C and C++ High Performance Linux Clusters with OSCAR, Rocks, OpenMosix, and MPI oreilly.com oreilly.com is more than a complete catalog of O’Reilly books. You’ll also find links to news, events, articles, weblogs, sample chapters, and code examples. oreillynet.com is the essential portal for developers interested in open and emerging technologies, including new platforms, pro- gramming languages, and operating systems. Conferences O’Reilly brings diverse innovators together to nurture the ideas that spark revolutionary industries. We specialize in document- ing the latest tools and systems, translating the innovator’s knowledge into useful skills for those in the trenches. Visit con- ferences.oreilly.com for our upcoming events. Safari Bookshelf (safari.oreilly.com) is the premier online refer- ence library for programmers and IT professionals. Conduct searches across more than 1,000 books. Subscribers can zero in on answers to time-critical questions in a matter of seconds. Read the books on your Bookshelf from cover to cover or sim- ply flip to the page you need. Try it today for free.

Intel Threading Building Blocks Outfitting C++ for Multi-Core Processor Parallelism James Reinders Beijing • Cambridge • Farnham • Köln • Paris • Sebastopol • Taipei • Tokyo

分享到：

赞收藏

资料库

Intel Threading Building Blocks.pdf

相关推荐

开发技术

热门标签

最新资料