BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260522T150122Z
LOCATION:C146
DTSTART;TZID=America/Chicago:20181114T143000
DTEND;TZID=America/Chicago:20181114T150000
UID:submissions.supercomputing.org_SC18_sess216_pap386@linklings.com
SUMMARY:Evaluating and Accelerating High-Fidelity Error Injection for HPC
DESCRIPTION:Chun-Kai Chang, Sangkug Lym, and Nicholas Kelly (University of
  Texas); Michael B. Sullivan (Nvidia Corporation); and Mattan Erez (Univer
 sity of Texas)\n\nWe address two important concerns in the analysis of the
  behavior of applications in the presence of hardware errors: (1) when is 
 it important to model how hardware faults lead to erroneous values (instru
 ction-level errors) with high fidelity, as opposed to using simple bit-fli
 pping models, and (2) how to enable fast high-fidelity error injection cam
 paigns, in particular when error detectors are employed. We present and ve
 rify a new nested Monte Carlo methodology for evaluating high-fidelity gat
 e-level fault models and error-detector coverage, which is orders of magni
 tude faster than current approaches. We use that methodology to demonstrat
 e that, without detectors, simple error models suffice for evaluating erro
 rs in 9 HPC benchmarks.\n\nTag: Performance, Resiliency, Tools\n\nRegistra
 tion Category: Tech Program Reg Pass\n\nSession Chair: Michèle Weiland (EP
 CC, The University of Edinburgh; The University of Edinburgh)\n\n
END:VEVENT
END:VCALENDAR