“GPU Accelerated Analytics in a Post Moore’s Law World” – Larry Brown Ph.D.
The use of GPUs for general purpose computing has grown significantly since NVIDIA released CUDA in 2007. In this presentation I will cover a brief history of GPU computing, describe how a GPU is different from a CPU and review the broad scope and rich ecosystem of GPGPU computing from embedded systems to the data center. I will highlight two recently emergent and exciting GPGPU workflows – one using GPUs for high-performance data analytics, and another where GPU accelerated deep learning is being applied to traditional HPC. In the end, I hope the audience will understand how GPUs and GPU powered AI allow engineers to “cheat” Moore’s law and capture dramatic gains in energy efficiency and compute performance.
Larry Brown is a Sr. Solutions Architect with NVIDIA and manager of the Federal Solutions Architecture team. His team helps customers design and deploy GPU accelerated workflows in high-performance computing, data analytics, and deep learning. Larry has 20 years of experience designing, implementing and supporting a variety of advanced software and hardware systems for defense and national security applications. Prior to joining NVIDIA, he worked for some technology start-ups, as well as large defense systems companies including General Dynamics, Textron, and Booz Allen. Larry has designed electro-optical systems for head-mounted displays and training simulators, developed GIS applications for multi-touch displays, and been a software engineer on UGV and UAV programs. He has a Ph.D. from the Johns Hopkins University in the area of Vision Science and a graduate certificate in Software Engineering from the University of Colorado.