GPU Programming with OpenACC

• An understanding of C/C++ or Fortran language and parallel programming concepts
• Basic knowledge of computer architecture and memory hierarchy
• Experience with command-line tools and code editors

Audience

• Developers who wish to learn how to use OpenACC to program heterogeneous devices and exploit their parallelism
• Developers who wish to write portable and scalable code that can run on different platforms and devices
• Programmers who wish to explore the high-level aspects of heterogeneous programming and optimize their code productivity

OpenACC is an open standard for heterogeneous programming that enables a code to run on different platforms and devices, such as multicore CPUs, GPUs, FPGAs, and others. 

This instructor-led, live training (online or onsite) is aimed at beginner-level to intermediate-level developers who wish to use OpenACC to program heterogeneous devices and exploit their parallelism.

By the end of this training, participants will be able to:

• Set up an OpenACC development environment.
• Write and run a basic OpenACC program.
• Annotate code with OpenACC directives and clauses.
• Use OpenACC API and libraries.
• Profile, debug, and optimize OpenACC programs.

Format of the Course

• Interactive lecture and discussion.
• Lots of exercises and practice.
• Hands-on implementation in a live-lab environment.

Course Customization Options

• To request a customized training for this course, please contact us to arrange.

Introduction

• What is OpenACC?
• OpenACC vs OpenCL vs CUDA vs SYCL
• Overview of OpenACC features and architecture
• Setting up the development environment

Getting Started

• Creating an OpenACC project in Visual Studio Code
• Exploring project structure and files
• Compiling and running the program
• Displaying output with printf and fprintf

OpenACC Directives and Clauses

• Understanding OpenACC directives and clauses
• Using parallel directives for creating parallel regions
• Using kernels directives for compiler-managed parallelism
• Using loop directives for parallelizing loops
• Managing data movement with data directives
• Synchronizing data with update directives
• Improving data reuse with cache directives
• Creating device functions with routine directives
• Synchronizing events with wait directives

OpenACC API

• Understanding the role of OpenACC API
• Querying device information and capabilities
• Setting device number and type
• Handling errors and exceptions
• Creating and synchronizing events

OpenACC Libraries and Interoperability

• Understanding OpenACC libraries and interoperability
• Using math, random, and complex libraries
• Integrating with other models (CUDA, OpenMP, MPI)
• Integrating with GPU libraries (cuBLAS, cuFFT)

OpenACC Tools

• Understanding OpenACC tools in development
• Profiling and debugging OpenACC programs
• Performance analysis with PGI Compiler, NVIDIA Nsight Systems, Allinea Forge

Optimization

• Factors affecting OpenACC program performance
• Optimizing data locality and reducing transfers
• Optimizing loop parallelism and fusion
• Optimizing kernel parallelism and fusion
• Optimizing vectorization and auto-tuning

Summary and Next Steps