OpenACC for Programmers

Sunita Chandrasekaran is assistant professor in the Computer and Information Sciences Department at the University of Delaware. Her research interests include exploring the suitability of high-level programming models and runtime systems for HPC and embedded platforms, and migrating scientific applications to heterogeneous computing systems. Dr. Chandrasekaran was a post-doctoral fellow at the University of Houston and holds a Ph.D. from Nanyang Technological University, Singapore. She is a member of OpenACC, OpenMP, MCA and SPEC HPG. She has served on the program committees of various conferences and workshops including SC, ISC, ICPP, CCGrid, Cluster, and PACT, and has co-chaired parallel programming workshops co-located with SC, ISC, IPDPS, and SIAM.Guido Juckeland is head of the Computational Science Group, Department for Information Services and Computing, Helmholtz-Zentrum Dresden-Rossendorf, and coordinates the work of the GPU Center of Excellence at Dresden. He and also represents HZDR at the SPEC High Performance Group and OpenACC committee. He received his Ph.D. from Technische Universität Dresden for his work on performance analysis for hardware accelerators. He was a Gordon Bell Award Finalist in 2013. Previously he worked as the IT-architect and post-doctoral researcher for the Center for Information Services and High Performance Computing (ZIH) at TU Dresden, Germany. He has served on the program committees of various conferences and workshops, including ISC, EuroPar, CCGrid, ASHES, P^3MA, PMBS, WACCPD, and PACT, and has co-chaired parallel programming workshops co-located with SC.

• Foreword xvPreface xxiAcknowledgments xxiiiAbout the Contributors xxvChapter 1: OpenACC in a Nutshell 11.1 OpenACC Syntax 31.2 Compute Constructs 61.3 The Data Environment 111.4 Summary 151.5 Exercises 15Chapter 2: Loop-Level Parallelism 172.1 Kernels Versus Parallel Loops 182.2 Three Levels of Parallelism 212.3 Other Loop Constructs 242.4 Summary 302.5 Exercises 31Chapter 3: Programming Tools for OpenACC 333.1 Common Characteristics of Architectures 343.2 Compiling OpenACC Code 353.3 Performance Analysis of OpenACC Applications 363.4 Identifying Bugs in OpenACC Programs 513.5 Summary 533.6 Exercises 54Chapter 4: Using OpenACC for Your First Program 594.1 Case Study 594.2 Creating a Naive Parallel Version 684.3 Performance of OpenACC Programs 714.4 An Optimized Parallel Version 734.5 Summary 784.6 Exercises 79Chapter 5: Compiling OpenACC 815.1 The Challenges of Parallelism 825.2 Restructuring Compilers 885.3 Compiling OpenACC 925.4 Summary 975.5 Exercises 97Chapter 6: Best Programming Practices 1016.1 General Guidelines 1026.2 Maximize On-Device Compute 1056.3 Optimize Data Locality 1086.4 A Representative Example 1126.5 Summary 1186.6 Exercises 119Chapter 7: OpenACC and Performance Portability 1217.1 Challenges 1217.2 Target Architectures 1237.3 OpenACC for Performance Portability 1247.4 Code Refactoring for Performance Portability1267.5 Summary 1327.6 Exercises133Chapter 8: Additional Approaches to Parallel Programming 1358.1 Programming Models1358.2 Programming Model Components1428.3 A Case Study 1558.4 Summary1708.5 Exercises170Chapter 9: OpenACC and Interoperability 1739.1 Calling Native Device Code from OpenACC 1749.2 Calling OpenACC from Native Device Code 1819.3 Advanced Interoperability Topics 1829.4 Summary1859.5 Exercises185Chapter 10: Advanced OpenACC 18710.1 Asynchronous Operations 18710.2 Multidevice Programming 20410.3 Summary 21310.4 Exercises 213Chapter 11: Innovative Research Ideas Using OpenACC, Part I 21511.1 Sunway OpenACC 21511.2 Compiler Transformation of Nested Loops for Accelerators 224Chapter 12: Innovative Research Ideas Using OpenACC, Part II 23712.1 A Framework for Directive-Based High-Performance Reconfigurable Computing 23712.2 Programming Accelerated Clusters Using XcalableACC 253Index 269