Jason Sanders, NVIDIA Friday, February 19, 4:30 PM Small Auditorium (CS 105) Abstract: The computing industry is currently undergoing a parallel computing revolution. The era of upward-spiraling clock speeds has drawn to a close, and the era of increasing core counts is upon us. The graphics processing unit (GPU) features a massively parallel computing architecture and is well poised to figure heavily in this evolving landscape. Two programming interfaces for the GPU, CUDA C and OpenCL, are readily available. In this presentation, NVIDIA GPU architecture and its programming models are explored. Bio: Jason Sanders is currently a senior software engineer in the CUDA team. He has been at NVIDIA for over two years, where he has developed system software and most recently has completed a book on CUDA C, due to hit shelves in 2010. He received his BSE from Princeton University in 2001 and an MS from the University of California Berkeley in 2007.