Introduction to OpenMP GPU Offloading

Introduction to OpenMP GPU Offloading
September 22-23, 2021

On September 22-23 from 1-3:30 PM (ET)  OLCF and NERSC
will offer a (virtual) Introduction to OpenMP GPU Offloading. This training is meant for OLCF
and NERSC users who are already familiar with the basic ideas of GPU programming but who want to learn
about the core GPU offloading capabilities of OpenMP.

During each day of the training, OLCF/NERSC  staff will give a lecture followed by hands-on exercises forthe participants to work through. On day 1, we will give a general overview of the OpenMP
programming model and cover the basics of using OpenMP directives to offload work to GPUs.
On day 2, we will discuss optimization strategies and show how efficient data movement and a
better understanding of the hierarchy of parallelism available can lead to improved
performance. The hands-on exercises will be available in both C and Fortran and will be
designed to work with a variety of compilers. Time will be reserved for users to work through
the examples during the training and a slack channel will be available for questions during and
after each training session.

OLCF users will use their Summit accounts and NERSC users will use Cori GPU.

Draft Agenda
All times given in EDT

September 22
1:00-2:00 pm Introductions to OpenMP Offloading
2-3:30 pm Hands on examples
3:30-5:00 pm Help available on slack

September 23
1:00-2:00 pm The Memory model and Efficient Data Transfer.
2-3:30 pm Hands on examples
3:30-5:00 pn Help available on Slack