Learn How GPUs Accelerate HPC

REGISTER

Massively parallel NVIDIA GPUs provide the bulk of the computing power behind many of the World's top supercomputers. This talk will start with the basic architecture of the GPU and introduce the CUDA parallel computing platform. The basic principles of massively parallel computing will be introduced using simple source code examples of CUDA C and C++. Learn about the latest software and tools, from where to download and install, to how to get started writing your own parallel code in Linux, Mac or Windows environments.

Attendees will learn:
• Fundaments of general purpose GPU Computing
• Concepts of massively parallel computing using CUDA
• How do get started with CUDA C/C++ in Linux, Mac and Windows

About the speaker:

As CUDA Educator at NVIDIA, Mark Ebersole teaches developers the benefit of GPU computing using the NVIDIA CUDA parallel computing platform and programming model, and the benefits of GPU computing. With more than ten years of experience as a systems programmer, Mark has spent much of his time at NVIDIA as a GPU systems diagnostics programmer in which he developed a tool to test, debug, validate, and verify GPUs from pre-emulation through bringup and into production. Before joining NVIDIA, he worked at IBM developing Linux drivers for the IBM iSeries server. Mark holds a BS degree in math and computer science from St. Cloud State University.

REGISTER

The recently introduced Kepler architecture offers a wealth of new features to assist programmers developing applications on the GPU and help them achieve unprecedented performance. While many of these features are leveraged transparently by the NVCC compiler and the CUDA software tool chain, programmers can support these tools by targeting the Kepler architecture with their software design. The goal of this presentation is to provide CUDA developers with an understanding of the key concepts of the Kepler architecture and demonstrate how they can be used in real world applications. After a review of the Kepler architecture and a brief introduction of general GPU optimization strategies, we will present an in-depth look at Kepler features targeting both coarse-grain and fine-grain parallelism, including dynamic parallelism, Hyper-Q, warp shuffle and more. This will provide CUDA developers with the necessary background to make the optimal design choices for Kepler.

What you'll learn:
• New Kepler features including CUDA Dynamic Parallelism, warp shuffle instructions
• How to use these features in real-world applications
• General GPU optimization techniques

About the speaker:

Peter Messmer joined NVIDIA in 2011 after spending more than 15 years developing HPC and GPU accelerated applications for industry and Government clients, mainly in the area of plasma and EM simulations, data analysis and visualization. In his role as Senior Devtech Engineer at NVIDIA, Peter is working with HPC users around the globe supporting them in accelerating their scientific discovery process by taking advantage of GPUs in their applications. Peter holds a PhD and MSc in Physics from ETH Zurich, Switzerland, with specialization in kinetic plasma physics and nonlinear optics.