IISWC-2014

October 26-28, 2014

 Raleigh, North Carolina, USA


KEYNOTE: Kevin Skadron (University of Virginia)

TITLE
Automata Processing, and the Path to a Multi-Accelerator Future  

 

ABSTRACT
Power and cooling constraints have brought multicore and accelerator architectures -- notably GPUs and a variety of mobile SoCs -- into widespread commercial use. The resulting speedups have increased interest in development of specialized accelerators, as well as a renewed interest in reconfigurable logic. However, the questions of how to incorporate such hardware into general-purpose computer systems, and how to develop effective programming models, are difficult, because adding accelerators complicates the design space for both hardware and software. This talk will briefly review the landscape of heterogeneous computing, and use Micron's recently-introduced Automata Processor as a case study in exploring some specific challenges and opportunities for future heterogeneous systems.  

 

BIO
Kevin Skadron is the founding director of the Center for Automata Processing, as well as professor and chair of computer science, at the University of Virginia, where he has been on the faculty since 1999. He received his Ph.D. in Computer Science from Princeton University, and his BS and BA degrees in Computer Engineering and Economics from Rice University. He is a Fellow of the IEEE and a recipient of the ACM SIGARCH Maurice Wilkes Award. Skadron's research interests focus on design and application of accelerators and heterogeneous architectures, including solutions to power, thermal, reliability, and programming challenges. To support research in these areas, he and his colleagues have developed the Rodinia benchmark suite for heterogeneous computing and contributed to the new SPEC ACCEL suite, as well as developing the HotSpot, VoltSpot, and ArchFP modeling tools. The recently-created Center for Automata Processing seeks to catalyze research on applications, programming models, and heterogeneous architectures leveraging automata-based accelerators.