Consumer-focused High Performance Computing Architectures
Abstract: The technology landscape is incredibly exciting today, with high-performance computation transforming many aspects of society and daily life. New innovations appear seemingly daily in areas of entertainment, transportation, communication, and health care, just to name a few. Emerging practical applications of virtual and augmented reality, autonomous vehicles, and automated reasoning will place new demands on our computing architectures. While the computational appetite of emerging applications in these spaces appear to be growing without bound, the historical technology scaling trends which have provided the fundamental horsepower for computing over the last 50 years, are slowing substantially. This talk will discuss some of the cataclysmic trends in consumer applications of high-performance computing, and focus on opportunities for computer designers. It will also present challenges associated with emerging deep neural networks (DNNs) and describe recent works that (1) enable larger and more complex networks to be trained on compute devices with limited memory capacity; and (2) reduce the memory and computation footprints of DNNs at inference time, enabling them to run with vastly improved energy efficiency.
Bio: Dr. Stephen W. Keckler is the Vice President of Architecture Research at NVIDIA and an Adjunct Professor of Computer Science at the University of Texas at Austin, where he served on the faculty from 1998-2012. His research interests include parallel computer architectures, high-performance computing, energy-efficient architectures, and embedded computing. Dr. Keckler is a Fellow of the ACM, a Fellow of the IEEE, an Alfred P. Sloan Research Fellow, and a recipient of the NSF CAREER award, the ACM Grace Murray Hopper award, the President’s Associates Teaching Excellence Award at UT-Austin, and the Edith and Peter O’Donnell award for Engineering. He earned a B.S. in Electrical Engineering from Stanford University and M.S. and Ph.D. degrees in Computer Science from the Massachusetts Institute of Technology.