book

OpenACC for Programmers: Concepts and Strategies, First Edition

Name: OpenACC for Programmers: Concepts and Strategies, First Edition
ISBN: 9780134694306

by Guido Juckeland, Sunita Chandrasekaran

September 2017

Intermediate to advanced

320 pages

9h 25m

English

Addison-Wesley Professional

Read now

Unlock full access

Title Page
Copyright Page
Dedication Page
Contents
Foreword
Preface
Acknowledgments
About the Contributors
Chapter 1: OpenACC in a Nutshell
1.1 OpenACC Syntax1.1.1 Directives1.1.2 Clauses1.1.3 API Routines and Environment Variables1.2 Compute Constructs1.2.1 Kernels1.2.2 Parallel1.2.3 Loop1.2.4 Routine1.3 The Data Environment1.3.1 Data Directives1.3.2 Data Clauses1.3.3 The Cache Directive1.3.4 Partial Data Transfers1.4 Summary1.5 Exercises
Chapter 2: Loop-Level Parallelism
2.1 Kernels Versus Parallel Loops2.2 Three Levels of Parallelism2.2.1 Gang, Worker, and Vector Clauses2.2.2 Mapping Parallelism to Hardware2.3 Other Loop Constructs2.3.1 Loop Collapse2.3.2 Independent Clause2.3.3 Seq and Auto Clauses2.3.4 Reduction Clause2.4 Summary2.5 Exercises

Chapter 3: Programming Tools for OpenACC
3.1 Common Characteristics of Architectures3.2 Compiling OpenACC Code3.3 Performance Analysis of OpenACC Applications3.3.1 Performance Analysis Layers and Terminology3.3.2 Performance Data Acquisition3.3.3 Performance Data Recording and Presentation3.3.4 The OpenACC Profiling Interface3.3.5 Performance Tools with OpenACC Support3.3.6 The NVIDIA Profiler3.3.7 The Score-P Tools Infrastructure for Hybrid Applications3.3.8 TAU Performance System3.4 Identifying Bugs in OpenACC Programs3.5 Summary3.6 Exercises
Chapter 4: Using OpenACC for Your First Program
4.1 Case Study4.1.1 Serial Code4.1.2 Compiling the Code4.2 Creating a Naive Parallel Version4.2.1 Find the Hot Spot4.2.2 Is It Safe to Use kernels?4.2.3 OpenACC Implementations4.3 Performance of OpenACC Programs4.4 An Optimized Parallel Version4.4.1 Reducing Data Movement4.4.2 Extra Clever Tweaks4.4.3 Final Result4.5 Summary4.6 Exercises
Chapter 5: Compiling OpenACC
5.1 The Challenges of Parallelism5.1.1 Parallel Hardware5.1.2 Mapping Loops5.1.3 Memory Hierarchy5.1.4 Reductions5.1.5 OpenACC for Parallelism5.2 Restructuring Compilers5.2.1 What Compilers Can Do5.2.2 What Compilers Can’t Do5.3 Compiling OpenACC5.3.1 Code Preparation5.3.2 Scheduling5.3.3 Serial Code5.3.4 User Errors5.4 Summary5.5 Exercises
Chapter 6: Best Programming Practices
6.1 General Guidelines6.1.1 Maximizing On-Device Computation6.1.2 Optimizing Data Locality6.2 Maximize On-Device Compute6.2.1 Atomic Operations6.2.2 Kernels and Parallel Constructs6.2.3 Runtime Tuning and the If Clause6.3 Optimize Data Locality6.3.1 Minimum Data Transfer6.3.2 Data Reuse and the Present Clause6.3.3 Unstructured Data Lifetimes6.3.4 Array Shaping6.4 A Representative Example6.4.1 Background: Thermodynamic Tables6.4.2 Baseline CPU Implementation6.4.3 Profiling6.4.4 Acceleration with OpenACC6.4.5 Optimized Data Locality6.4.6 Performance Study6.5 Summary6.6 Exercises
Chapter 7: OpenACC and Performance Portability
7.1 Challenges7.2 Target Architectures7.2.1 Compiling for Specific Platforms7.2.2 x86_64 Multicore and NVIDIA7.3 OpenACC for Performance Portability7.3.1 The OpenACC Memory Model7.3.2 Memory Architectures7.3.3 Code Generation7.3.4 Data Layout for Performance Portability7.4 Code Refactoring for Performance Portability7.4.1 HACCMK7.4.2 Targeting Multiple Architectures7.4.3 OpenACC over NVIDIA K20x GPU7.4.4 OpenACC over AMD Bulldozer Multicore7.5 Summary7.6 Exercises
Chapter 8: Additional Approaches to Parallel Programming
8.1 Programming Models8.1.1 OpenACC8.1.2 OpenMP8.1.3 CUDA8.1.4 OpenCL8.1.5 C++ AMP8.1.6 Kokkos8.1.7 RAJA8.1.8 Threading Building Blocks8.1.9 C++178.1.10 Fortran8.2 Programming Model Components8.2.1 Parallel Loops8.2.2 Parallel Reductions8.2.3 Tightly Nested Loops8.2.4 Hierarchical Parallelism (Non-Tightly Nested Loops)8.2.5 Task Parallelism8.2.6 Data Allocation8.2.7 Data Transfers8.3 A Case Study8.3.1 Serial Implementation8.3.2 The OpenACC Implementation8.3.3 The OpenMP Implementation8.3.4 The CUDA Implementation8.3.5 The Kokkos Implementation8.3.6 The TBB Implementation8.3.7 Some Performance Numbers8.4 Summary8.5 Exercises
Chapter 9: OpenACC and Interoperability
9.1 Calling Native Device Code from OpenACC9.1.1 Example: Image Filtering Using DFTs9.1.2 The host_data Directive and the use_device Clause9.1.3 API Routines for Target Platforms9.2 Calling OpenACC from Native Device Code9.3 Advanced Interoperability Topics9.3.1 acc_map_data9.3.2 Calling CUDA Device Routines from OpenACC Kernels9.4 Summary9.5 Exercises
Chapter 10: Advanced OpenACC
10.1 Asynchronous Operations10.1.1 Asynchronous OpenACC Programming10.1.2 Software Pipelining10.2 Multidevice Programming10.2.1 Multidevice Pipeline10.2.2 OpenACC and MPI10.3 Summary10.4 Exercises
Chapter 11: Innovative Research Ideas Using OpenACC, Part I
11.1 Sunway OpenACC11.1.1 The SW26010 Manycore Processor11.1.2 The Memory Model in the Sunway TaihuLight11.1.3 The Execution Model11.1.4 Data Management11.1.5 Summary11.2 Compiler Transformation of Nested Loops for Accelerators11.2.1 The OpenUH Compiler Infrastructure11.2.2 Loop-Scheduling Transformation11.2.3 Performance Evaluation of Loop Scheduling11.2.4 Other Research Topics in OpenUH
Chapter 12: Innovative Research Ideas Using OpenACC, Part II
12.1 A Framework for Directive-Based High-Performance Reconfigurable Computing12.1.1 Introduction12.1.2 Baseline Translation of OpenACC-to-FPGA12.1.3 OpenACC Extensions and Optimization for Efficient FPGA Programming12.1.4 Evaluation12.1.5 Summary12.2 Programming Accelerated Clusters Using XcalableACC12.2.1 Introduction to XcalableMP12.2.2 XcalableACC: XcalableMP Meets OpenACC12.2.3 Omni Compiler Implementation12.2.4 Performance Evaluation on HA-PACS12.2.5 Summary
Index
Register Your Product

Overview

The Complete Guide to OpenACC for Massively Parallel Programming

Scientists and technical professionals can use OpenACC to leverage the immense power of modern GPUs without the complexity traditionally associated with programming them. OpenACC™ for Programmers is one of the first comprehensive and practical overviews of OpenACC for massively parallel programming.

This book integrates contributions from 19 leading parallel-programming experts from academia, public research organizations, and industry. The authors and editors explain each key concept behind OpenACC, demonstrate how to use essential OpenACC development tools, and thoroughly explore each OpenACC feature set.

Throughout, you’ll find realistic examples, hands-on exercises, and case studies showcasing the efficient use of OpenACC language constructs. You’ll discover how OpenACC’s language constructs can be translated to maximize application performance, and how its standard interface can target multiple platforms via widely used programming languages.

Each chapter builds on what you’ve already learned, helping you build practical mastery one step at a time, whether you’re a GPU programmer, scientist, engineer, or student. All example code and exercise solutions are available for download at GitHub.

Discover how OpenACC makes scalable parallel programming easier and more practical
Walk through the OpenACC spec and learn how OpenACC directive syntax is structured
Get productive with OpenACC code editors, compilers, debuggers, and performance analysis tools
Build your first real-world OpenACC programs
Exploit loop-level parallelism in OpenACC, understand the levels of parallelism available, and maximize accuracy or performance
Learn how OpenACC programs are compiled
Master OpenACC programming best practices
Overcome common performance, portability, and interoperability challenges
Efficiently distribute tasks across multiple processors

Register your product at informit.com/register for convenient access to downloads, updates, and/or corrections as they become available.

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.

Read now

Unlock full access

More than 5,000 organizations count on O’Reilly

O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.

Julian F.

Head of Cybersecurity

I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.

Addison B.

Field Engineer

I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.

Amir M.

Data Platform Tech Lead

I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.

Mark W.

Embedded Software Engineer

Publisher Resources

ISBN: 9780134694306

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills