Description

This course gives a thorough introduction to programming GPUs using the directive based OpenACC paradigm. The course consists of lectures and hands-on exercises. Topics of this course include the basic usage of OpenACC, as well as some more advanced issues related to profiling, performance and interoperability with CUDA and MPI.

Learning outcome

After the course the participants should have the basic skills needed for utilizing OpenACC in new, or existing programs.

Prerequisites

The participants are assumed to have working knowledge of Fortran and/or C programming languages. In addition, fluent operation in a Linux/Unix environment will be assumed.

Agenda

Day 1, Monday 14.10

09:00 - 12:00 SESSION 1 & Coffee break (10:00 -10:15)

Introduction to accelerators​
Introduction to OpenACC
Exercises

12:00 - 13:00 Lunch

13:00 - 16:00 SESSION 2 & Coffee break (14:15-14:30)

Data movement
Exercises

Day 2, Tuesday 15.10

09:00 - 12:00 SESSION 3 & Coffee break (10:15-10:30)

Profilling
Performance considerations
Exercises

12:00 - 13:00 Lunch

13:00 - 16:00 SESSION 4 & Coffee break (14:00-14:15)

Asynchronous operations and pipelining
Interoperability with CUDA and GPU-Accelerated libraries
Exercises

Lecturers: 

Sebastian von Alfthan (CSC), Fredrick Robertsén (CSC)

Language:  English
Price:          Free of charge
https://events.prace-ri.eu/event/887/