2011-04-14 Keeneland Workshop
Workshop Tutorial
The tutorial presentations can be found at:
On April 14-15, 2011, the Keeneland project and the Georgia Tech NVIDIA CUDA Center of Excellence presented a two day tutorial on GPU heterogeneous processing for computational science.
Tutorial topics included
- Architectural overview and motivation
- Introduction to CUDA
- Introduction to OpenCL
- Application case studies
- Overviews of GPU enabled libraries, correctness and performance tools, and compilers
- High level language interfaces, such as MATLAB, R, and Accelereyes, to GPUs
- Advanced optimization techniques for GPUs
- Hands-on time with the Keeneland Initial Delivery System on Friday (4/15) afternoon
Agenda
Thursday 4/14
Time/ Topic
08:00 Welcome and Introduction
08:15 Heterogeneous Architectures and KID Overview
08:45 Introduction to CUDA
10:15 Break
10:30 Introduction to OpenCL
11:30 Lunch
12:30 SDK Overview, incl debugging and performance
13:00 Libraries Overview: NVIDIA Libraries
13:40 Libraries Overview: MAGMA
14:00 Scalable Heterogeneous Computing
14:45 Break
15:00 Integrated Kernal Walk-throughs
15:45 Ocelot
16:15 Performance Analysis with Tau
17:00 Adjourn
Friday 4/15
08:00 Announcements and Recap
08:05 Allinea DDT
08:35 Advanced Topics in CUDA, OpenCL
09:35 break
9:50 Additional Tools, Compilers, Libraries
10:30 Application Case Studies
11:30 Lunch
12:30 HANDS ON WITH KEENELAND - Bring your own code, use SHOC, or NVIDIA SDK
17:00 Adjourn
The Keeneland Workshop is sponsored by the Georgia Institute of Technology NVIDIA CUDA Center of Excellence and the Keeneland Project. The Keeneland Project is funded by the National Science Foundation.