home · mobile · calendar · colloquia · 2004-2005 · 

Colloquium - Owens

High-Performance Data-Parallel Computing on the Desktop with the GPU
University of California, Davis

With the recent announcements by Intel and AMD that mainline future microprocessors will feature multiple cores, and the debut of IBM's new Cell processor, it is clear that the desktop computers of tomorrow will embrace parallelism in both hardware and software. Yet the first commodity parallel processor is already on the desktops of most users today: the graphics processor. In this talk I will discuss the evolution of graphics hardware from fixed-function, special-purpose processors to programmable, data-parallel Graphics Processing Units (GPUs). I will concentrate the discussion on the use of these processors for general-purpose computation (GPGPU) in such diverse areas as signal and image processing, linear algebra and numerical methods, database operations, and geometric computing, and what this may mean for future computing.

John Owens is an assistant professor of electrical and computer engineering at the University of California at Davis. At Davis, he is affiliated with the Institute for Data Analysis and Visualization and leads research groups in projects in graphics hardware and sensor networks. He graduated from Stanford University in 2003 with a PhD in electrical engineering and the University of California, Berkeley, in 1995 with a BS in electrical engineering and computer sciences.

Department of Computer Science
University of Colorado Boulder
Boulder, CO 80309-0430 USA
May 5, 2012 (14:13)