Tutorial Details: NVIDIA CUDA Talk
|Date:||Wednesday, June 13, 2012, 09:00 am - 02:00 pm|
This tutorial is presented by Mark Ebersole, NVIDIA CUDA Developer.
9:00-10:00 Introduction to GPU Computing; CUDA 101 - An introduction to what GPU Computing is about and a brief tutorial on programming with CUDA C.
10:30-11:30 CUDA 5 and Beyond - Come learn what's new and exciting with CUDA 5 and the Kepler architecture, and how it can further accelerate your science. Finally, where do we see CUDA going?
12:00-1:00 Solving Sparse & Irregular Problems on GPUs
Modern GPUs are widely recognized to excel at dense, regular computations rich with floating point arithmetic. Their throughput-oriented architecture provides tremendous raw potential, and dense computations typically achieve high percentages of peak performance. What is less widely recognized is that GPUs also perform very well on sparse, irregular computations. In fact, their high memory bandwidth coupled with massive multithreading for latency tolerance makes them well suited to these kinds of computations as well.
Event info: http://www.facebook.com/events/323183474424250