processors

Talk Abstract:
Many aspects inherent to continued deep

-submicron scaling collude to impair the reliability of future microprocess

or implementations. This talk develops the idea of fingerprinting as an imp

ortant ingredient for efficient error detection. A fingerprint is a hashed

signature of internal state changes of a digital system. For example when a

pplied at the architectural level one may compute the fingerprint of the r

egister file and/or cache updates. For the purpose of detecting differences
in the mirrored operation of two processors comparing their fingerprints

for agreement is nearly as effective as the daunting alternative of compari

ng instantaneously all their internal states. We present two applications o

f fingerprinting. The first employs architectural fingerprinting to support
dual-modular-redundant execution in a multi-core processor. Fingerprinting
and other techniques combine to enable two mirrored cores to maintain redu

ndant execution and checking without requiring them to be microarchitectura

lly deterministic or to be in precise locked-step. The second work applies

microarchitectural-level fingerprinting to extremely-high-coverage detectio

n of transient failures in the datapath that would normally be masked and g

one unnoticed at the architectural level. This capability is central to our
approach to preemptively detect the on-set of transistor wear-out failures

. This talk presents joint work with Prof Babak Falsafi in the TRUSS projec

t (http://www.ece.cmu.edu/%7Etruss/) at the Computer Architecture Lab at Ca

rnegie Mellon (CALCM).

Speaker Bio:
James C. Hoe is an Associate

Professor of Electrical and Computer Engineering at Carnegie Mellon Univers

ity. His research interests include many aspects of computer architecture a

nd digital hardware design. His current research develops architecture and

microarchitecture solutions to improve computer reliability. His is also wo

rking on a hardware synthesis tool that compiles formal mathematical specif

ication of linear DSP transforms to hardware implementations. He received t

he B.S. degree in electrical engineering and computer science from Universi

ty of California at Berkeley in 1992 and the M.S. and Ph.D. degrees in elec

trical engineering and computer science from Massachusetts Institute of Tec

hnology in 1994 and 2000 respectively. For more information please visit