并行开发艺术(影印版) 下载 pdf 百度网盘 epub 免费 2025 电子版 mobi 在线

如果你想利用并发程序设计充分发挥多核心处理器的性能，《并行开发艺术(影印版)》就为你提供了所需要的实践知识和亲身体验。《并行开发艺术》是一份难得的材料，专注于多核心处理器的共享内存模型，而不只是理论模型或分布式内存架构。《并行开发艺术(影印版)》提供了详尽的解释和可用的示例，帮助你将算法从串行代码转化为并行代码，此外还包括建议和分析，避免程序员的典型错误。

作者为Intel资深工程师，拥有超过20年的并行和并发编程经验，《并行开发艺术(影印版)》将会帮助你：

· 探索共享内存与分布式内存编程间的区别

· 学习设计多线程程序的指导方针，包括测试和调整

· 研究如何善用不同的线程库，包括Windows线程、POSIX线程、OpenMP和Intel Threading Building Blocks

· 研究如何实现排序、查找、图形和其他实用计算的并行算法

《并行开发艺术》向你展示如何扩展算法，以从新型的多核处理器中获益。对于开发并行算法和并发编程来说，《并行开发艺术(影印版)》不可或缺。

PREFACE

1 WANT TO GO FASTER? RAISE YOUR HANDS IF YOU WANT TO GO FASTER!

Some Questions You May Have

Four Steps of a Threading Methodology

Background of Parallel Algorithms

Shared-Memory Programming Versus Distributed-Memory Programming

This Book’s Approach to Concurrent Programming

2 CONCURRENT OR NOT CONCURRENT?

Design Models for Concurrent Algorithms

What’s Not Parallel

3 PROVING CORRECTNESS AND MEASURING PERFORMANCE

Verification of Parallel Algorithms

Example: The Critical Section Problem

Performance Metrics (How Am I Doing?)

Review of the Evolution for Supporting Parallelism in Hardware

4 EIGHT SIMPLE RULES FOR DESIGNING MULTITHREADED APPLICATIONS

Rule 1: Identify Truly Independent Computations

Rule 2: Implement Concurrency at the Highest Level Possible

Rule 3: Plan Early for Scalability to Take Advantage of Increasing Numbers of Cores

Rule 4: Make Use of Thread-Safe Libraries Wherever Possible

Rule 5: Use the Right Threading Model

Rule 6: Never Assume a Particular Order of Execution

Rule 7: Use Thread-Local Storage Whenever Possible or Associate Locks to Specific Data

Rule 8: Dare to Change the Algorithm for a Better Chance of Concurrency

Summary

5 THREADING LIBRARIES

Implicit Threading

Explicit Threading

What Else Is Out There?

Domain-Specific Libraries

6 PARALLEL SUM AND PREFIX SCAN

Parallel Sum

Prefix Scan

Selection

A Final Thought

7 MAPREDUCE

Map As a Concurrent Operation

Reduce As a Concurrent Operation

Applying MapReduce

MapReduce As Generic Concurrency

8 SORTING

Bubblesort

Odd-Even Transposition Sort

Shellsort

Quicksort

Radix Sort

9 SEARCHING

Unsorted Sequence

Binary Search

10 GRAPH ALGORITHMS

Depth-First Search

All-Pairs Shortest Path

Minimum Spanning Tree

11 THREADING TOOLS

Debuggers

Performance Tools

Anything Else Out There?

Go Forth and Conquer

GLOSSARY

PHOTO CREDITS

INDEX

Clay Breshears博士，是Intel公司的课程架构师，专攻多核心及多线程程序设计与培训。

暂无出版社相关信息，正在全力查找中！

Two types of dependencies can occur between tasks. The first is order dependency, where sometask relies on the completed results of the computations from another task. This reliance canbe a direct need to use the computed values as input to the succeeding task, or it may simplybe the case that the task that follows will be updating the same memory locations as theprevious task and you must ensure that all of the previous updates have been completed beforeproceeding. Both of these cases describe a potential data race, which we need to avoid.For example, if you are building a house, putting up the roof involves attaching the rafters tothe walls, laying down the decking, and applying the shingles. The dependence between thesethree tasks is one of execution order. You can't put down shingles until the decking is there,and you can't nail down the decking unless you have the rafters in place. So, instead of hiringthree teams to do these three tasks in parallel, you can hire one roofing crew to do all three inthe order required （there is parallelism within each of the roofing steps, plus the act of puttingon the roof is independent of installing the electrical wiring, the plumbing, and putting updrywall）.

To satisfy an execution order constraint, you can schedule tasks that have an order dependencyonto the same thread and ensure that the thread executes the tasks in the proper sequence.The serial code was written with the order dependency already taken care of. So, the serialalgorithm should guide the correct decomposition of the computations into tasks andassignment of those tasks to threads. Still, even after grouping tasks to execute on threads,there may be order constraints between threads. If regrouping tasks to threads is not an optionor will severely hurt performance, you will be forced to insert some form of synchronizationto ensure correct execution order.

The second type of dependency is data dependency. Identifying potential data dependenciescan be straightforward: look for those variables that are featured on the left side of theassignment operator. Data races require that the variable in question have at least one threadthat is writing to that variable. Check for any assignment of values to the same variable thatmight be done concurrently as well as any updates to a variable that could be read concurrently.Of course, using pointers to reference memory locations can make the identification processtrickier. There are tools （covered in Chapter 11） that can assist in finding nonobvious datadependencies in your code.

在线阅读地址：并行开发艺术(影印版)在线阅读

在线听书地址：并行开发艺术(影印版)在线收听

在线购买地址：并行开发艺术(影印版)在线购买

暂无原文赏析，正在全力查找中！

编辑推荐

《并行开发艺术(影印版)》是由东南大学出版社出版的。

媒体评论

“这本书紧扣题目，令人赏心悦目。艺术无法传授，但追随大师的脚印便可觅得踪迹。作者在超级计算机公司工作30余年，而后又在学术界浸淫10年，我可以毫无疑问地说，本书实至名归。”

　　──Tom Murphy，Contra Costa学院，计算机科学程序委员会主席

“终于，一本专注于

前言

.