BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Colloquium : Disciplined Approximate Computing: From Language t
 o Hardware
DTSTART:20130125T161500
DTEND:20130125T173000
DTSTAMP:20260407T065848Z
UID:2c59a7468332e87924a4df1e9c988957eb5d735f9ca640df39e1131d
CATEGORIES:Conferences - Seminars
DESCRIPTION:Luis Ceze\, University of Washington\nAbstract\nEnergy is incr
 easingly a first-order concern in computer systems. Exploiting energy-accu
 racy trade-offs is an attractive choice in applications that can tolerate 
 inaccuracies. Recent work has explored exposing this trade-off in programm
 ing models. A key challenge\, though\, is how to isolate parts of the prog
 ram that must be precise from those that can be approximated so that a pro
 gram functions correctly even as quality of service degrades.\nIn this tal
 k I will described our effort on co-designing language and hardware to tak
 e advantage of approximate computing for significant energy savings. We us
 e type qualifiers to declare data that may be subject to approximate compu
 tation. Using these types\, the system automatically maps approximate vari
 ables to low-power storage\, uses low-power operations\, and even applies 
 more energy-efficient algorithms provided by the programmer. In addition\,
  the system can statically guarantee isolation of the precise program comp
 onent from the approximate component. This allows a programmer to control 
 explicitly how information flows from approximate data to precise data. Im
 portantly\, employing static analysis eliminates the need for dynamic chec
 ks\, further improving energy savings. I will describe a micro-architectur
 e that offers explicit approximate storage and computation and our recent 
 proposal on using neural networks as approximate accelerators for imperati
 ve programs. I will conclude with an overview of our current/future resear
 ch directions\, including language extensions for quality-of-result specif
 ication\, programming tools\,\napproximate storage and approximate wireles
 s communication.Biography\nLuis Ceze is an Associate Professor in the Comp
 uter Science and Engineering Department at the University of Washington. 
  His research focuses on computer architecture\, programming languages and
  OS to improve the programmability\, reliability and energy efficiency of 
 multiprocessor systems. He has co-authored over 60 papers in these areas\,
  and had several papers selected as IEEE Micro Top Picks and CACM research
  Highlights. He participated in the Blue Gene\, Cyclops\,\nand PERCS proje
 cts at IBM and is a recipient of IBM awards. He is also a recipient of an 
 NSF CAREER Award\, a Sloan Research Fellowship and a Microsoft Research Fa
 culty Fellowship. He consults for Microsoft Research and co-founded Corens
 ic\, a UW-CSE spin-off company. He was born and raised in Sao Paulo\, Braz
 il\, where it drizzles all the time\; he now lives in the similarly drizzl
 y Seattle. When he is not working he is found either eating or cooking.
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR