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:C141/143/149
DTSTART;TZID=America/Chicago:20181113T113000
DTEND;TZID=America/Chicago:20181113T120000
UID:submissions.supercomputing.org_SC18_sess203_pap111@linklings.com
SUMMARY:Doomsday: Predicting Which Node Will Fail When on Supercomputers
DESCRIPTION:Anwesha Das and Frank Mueller (North Carolina State University
 ) and Paul Hargrove, Eric Roman, and Scott Baden (Lawrence Berkeley Nation
 al Laboratory)\n\nPredicting which node will fail and how soon remains a c
 hallenge for HPC resilience, yet may pave the way to exploiting proactive 
 remedies before jobs fail. Not only for increasing scalability up to exasc
 ale systems, but even for contemporary supercomputer architectures does it
  require substantial efforts to distill anomalous events from noisy raw lo
 gs. To this end, we propose a novel phrase extraction mechanism called TBP
  (time-based phrases) to pin-point node failures, which is unprecedented. 
  Our study, based on real system data and statistical machine learning, de
 monstrates the feasibility to predict which specific node will fail in Cra
 y systems. TBP achieves no less than 83% recall rates with lead times as h
 igh as 2 minutes. This opens up the door for enhancing prediction lead tim
 es for supercomputing systems in general, thereby facilitating efficient u
 sage of both computing capacity and power in large scale production system
 s.\n\nTag: GPUs, Resiliency, State of the Practice, System Software\n\nReg
 istration Category: Tech Program Reg Pass\n\nFinalist: BSP Finalist\n\nSes
 sion Chair: Ron Brightwell (Sandia National Laboratories)\n\n
END:VEVENT
END:VCALENDAR