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:20260522T150116Z
LOCATION:D174
DTSTART;TZID=America/Chicago:20181116T092000
DTEND;TZID=America/Chicago:20181116T094000
UID:submissions.supercomputing.org_SC18_sess146_ws_ftxs104@linklings.com
SUMMARY:Improving Application Resilience by Extending Error Correction wit
 h Contextual Information
DESCRIPTION:Alexandra L. Poulos and Dylan E. Wallace (Coastal Carolina Uni
 versity); Robert Robey, Laura Monroe, Vanessa Job, and Sean Blanchard (Los
  Alamos National Laboratory); William M. Jones (Coastal Carolina Universit
 y); and Nathan A. DeBardeleben (Los Alamos National Laboratory)\n\nExtreme
 -scale systems are growing in scope and complexity as we approach exascale
 . Uncorrectable faults in such systems are also increasing, so resilience 
 efforts addressing these are of great importance. In this paper, we extend
  a method that augments hardware error detection and correction (EDAC) con
 textually, and show an application-based approach that takes detectable un
 correctable (DUE) data errors and corrects them.\n\nWe applied this applic
 ation-based method successfully to data errors found using common EDAC, an
 d discuss operating system changes that will make this possible on existin
 g systems. We show that even when there are many acceptable correction cho
 ices (which may be seen in floating point), a large percentage of DUEs are
  corrected, and even the miscorrected data are very close to correct. We d
 eveloped two different contextual criteria for this application: local ave
 raging and global conservation of mass. Both did well in terms of closenes
 s, but conservation of mass outperformed averaging in terms of actual corr
 ectness.\n\nThe contributions of this paper are: 1) the idea of applicatio
 n-specific EDAC-based contextual correction, 2) its demonstration with gre
 at success on a real application, 3) the development of two different cont
 extual criteria, and 4) a discussion of attainable changes to the OS kerne
 l that make this possible on a real system.\n\nTag: Resiliency, Scientific
  Computing\n\nRegistration Category: Workshop Reg Pass\n\n
END:VEVENT
END:VCALENDAR