GPU Programming @ JSC

Apr 7, 2014, 9:00 AM → Apr 9, 2014, 4:30 PM CET

Computer Lab no. 021 (Jülich Supercomputing Centre)

Many-core programming is a very dynamic research area. Many scientific applications have been ported to GPU architectures in recent years. We will give an introduction to CUDA, OpenACC, OpenCL, and multi-GPU programming using examples of increasing complexity. After introducing the basics the focus will be on optimization and tuning of scientific applications. Topics covered will include:

• Programming models: CUDA, OpenACC, OpenCL
• Using libraries as interface for GPU programming (e.g. Thrust)
• Partitioning and granularity of parallel applications
• Debugging and profiling of kernels
• Performance optimizations
• Multi-GPU programming

Prerequisites: Knowledge in C

This course is a Prace Advanced Training Centres (PATC) course.

Application
Registrations are only considered until 7 March 2014;
due to available space, the maximal number of participants is limited.
Applicants will be notified, whether they are accepted for participitation.

Instructors:

Dr. Jan Meinke, Jochen Kreutz, Peter Philippen, Andrew Adinetz, Anke Zitz, JSC;
Jiri Kraus, NVIDIA

Contact
For any questions concerning the course please send an e-mail to w.homberg@fz-juelich.de

Monday, April 7

9:00 AM → 9:45 AM Introduction to GPU computing

9:45 AM → 10:30 AM Hands-on session

10:30 AM → 10:45 AM Coffee break

10:45 AM → 11:30 AM OpenACC intro

11:30 AM → 12:00 PM Hands-on session

12:00 PM → 1:00 PM Lunch Break

1:00 PM → 1:45 PM Advanced Features in OpenACC

1:45 PM → 2:30 PM Hands-on session

2:30 PM → 2:45 PM Coffe break

2:45 PM → 3:30 PM Cuda introduction

3:30 PM → 4:30 PM Hands-on session

7:00 PM → 10:00 PM Social event (Dinner at the Restaurant "Am Hexenturm" in Jülich)

Dinner at the Restaurant "Am Hexenturm" in Jülich

Tuesday, April 8

9:00 AM → 9:45 AM Matrix multiply using Cuda

9:45 AM → 10:30 AM Hands-on session

10:30 AM → 10:45 AM Coffee break

10:45 AM → 11:30 AM Cuda performance measurement

11:30 AM → 12:00 PM Hands-on session

12:00 PM → 1:00 PM Lunch Break

1:00 PM → 1:45 PM Performance tuning I

1:45 PM → 2:30 PM Hands-on session

2:30 PM → 2:45 PM Coffee break

2:45 PM → 3:30 PM Performance tuning II

3:30 PM → 4:30 PM Hands-on session

Wednesday, April 9

9:00 AM → 9:45 AM Intra-node multi GPU programming

9:45 AM → 10:30 AM Hands-on session

10:30 AM → 10:45 AM Coffee break

10:45 AM → 11:30 AM Multi GPU programming using Cuda/MPI

11:30 AM → 12:00 PM Hands-on session

12:00 PM → 1:00 PM Lunch break

1:00 PM → 1:45 PM OpenCL Overview

1:45 PM → 2:30 PM Hands-on session

2:30 PM → 2:45 PM Coffee break

2:45 PM → 3:30 PM Tuning OpenCL programs

3:30 PM → 4:15 PM Hands-on session

4:15 PM → 4:30 PM Wrap-up