Fundamentals of Accelerated Computing with OpenACC

Thursday Sep 16, 2021, 9:00 AM → 5:05 PM Europe/Berlin

Together with Nvidia, we are offering a workshop on accelerated computing with OpenACC.

The Workshop will take place virtually. Participants will be provided with access to a suitable GPU setup.

Connection details for the workshop as well as required steps for access to the setup will be sent to the participants prior to the event.

Learning Objectives

By participating in this workshop, you’ll have an understanding of the fundamental tools and techniques for GPU-accelerating C/C++ or Fortran applications with OpenACC and be able to:

• Profile and optimize your CPU-only applications to identify hot spots for acceleration.
• Use OpenACC directives to GPU-accelerate your codebase.
• Optimize data movement between the CPU and GPU accelerator.

Upon completion, you'll be ready to use OpenACC to GPU accelerate CPU-only applications.

Prerequisites

• Basic C/C++ or Fortran competency, including familiarity with variable types, loops, conditional statements, functions, and array manipulations.
• No previous knowledge of GPU programming is assumed.

 

This workshop is part of the Nvidia Deep Learning Institute.

 

1 Introduction

2 Introduction to Parallel Programming

10:05 AM Break

3 Profiling with OpenACC

Learn how to build and compile an OpenACC code, the importance of
profiling, and how to use the NVIDIA Nsight Systems profiler. Topics that will
be covered are as follows:

Compiling sequential and OpenACC code
The importance of code profiling
Profiling sequential and OpenACC multicore code
Technical introduction to the code used in introductory modules

11:15 AM Break

4 Introduction to OpenACC Directives

Learn how to parallelize your code with OpenACC directives and understand
the differences between parallel, kernel, and loop directives. Topics that will
be covered are as follows:

The Parallel directive
The Kernels directive
The Loop directive

12:50 PM Break

5 GPU Programming with OpenACC

Learn about the differences between GPUs and multicore CPUs, and manage
memory with CUDA Unified Memory. Topics that will be covered are as
follows:

Definition of a GPU
Basic OpenACC data management
CUDA Unified Memory
Profiling GPU applications

1:50 PM Break

6 Data Management with OpenACC

Learn how to explicitly manage data movement with OpenACC data directives
to reduce data transfers. Topics that will be covered are as follows:

OpenACC data directive/clauses
OpenACC structured data region
OpenACC unstructured data region
OpenACC update directive
Data management with C/C++ Structs/Classes

2:55 PM Break

7 Loop Optimizations with OpenACC

Understand the various levels of parallelism on a GPU and learn about ways
to extract more parallelism with OpenACC by optimizing loops in your code.
(50 mins) Topics that will be covered are as follows:

Seq/Auto clause
Independent clause
Reduction clause
Collapse clause
Tile clause
Gang, Worker, Vector

3:55 PM Break

8 Final Review

Review key learnings and answer questions.
Complete the assessment and earn a certificate.
Complete the workshop survey.