Heterogeneous Programming on GPUs with MPI + OmpSs

Venue: Barcelona Supercomputing Center. Barcelona, SPAIN.
Event date: 09 May 2018
Speaker(s): Xavier Martorell & Xavier Teruel
 
The tutorial will motivate the audience on the need for portable, efficient programming models that put less pressure on program developers while still getting good performance for clusters and clusters with GPUs.
 

Contents

More specifically, the tutorial will introduce the hybrid MPI/OmpSs parallel programming model for future exascale systems and it will demonstrate how to use MPI/OmpSs to incrementally parallelize/optimize: 1) MPI applications on clusters of SMPs, and 2) Leverage CUDA kernels with OmpSs on clusters of GPUs
 

Prerequisites

Good knowledge of C/C++
Basic knowledge of CUDA/OpenCL
Basic knowledge of Paraver/Extrae
 

Learning Outcomes

The students who finish this course will be able to develop benchmarks and simple applications with the MPI/OmpSs programming model to be executed in clusters of GPUs.
 

Agenda

Day 1
09.00h – Introduction to OmpSs
11.30h – OmpSs single node programming hands-on
13.00h – Lunch Break
14.00h – More on OmpSs: GPU/CUDA programming
15.00h – OmpSs single node programming hands-on with GPUs
17.30h – Adjourn
Day 2
09.00h – Introduction to MPI/OmpSs
10.00h – MPI/OmpSs hands-on
13.00h – Lunch Break
14.00h – Free hands-on session
17.30h – Adjourn
 

External links