BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20210402T160553Z
LOCATION:Track 11
DTSTART;TZID=America/New_York:20201111T115500
DTEND;TZID=America/New_York:20201111T122500
UID:submissions.supercomputing.org_SC20_sess204_ws_ftxs105@linklings.com
SUMMARY:From Tasks Graphs to Asynchronous Distributed Checkpointing with L
 ocal Restart
DESCRIPTION:Workshop\n\nFrom Tasks Graphs to Asynchronous Distributed Chec
 kpointing with Local Restart\n\nLion, Thibault\n\nThe ever-increasing numb
 er of computation units assembled in current HPC platforms leads to a conc
 erning increase in fault probability. Traditional checkpoint/restart strat
 egies avoid wasting large amounts of computation time when such fault occu
 rs. With the increasing amount of data processed by today's applications, 
 these strategies, however, suffer from their data transfer demand becoming
  unreasonable, or the entailed global synchronizations.\n\nThe current tre
 nd towards task-based programming is an opportunity to revisit the princip
 les of the checkpoint/restart strategies. We propose a checkpointing schem
 e which is closely tied to the execution of task graphs. We describe how i
 t allows for completely asynchronous and distributed checkpointing, as wel
 l as localized node restart, thus allowing for very large scalability. We 
 also show how a synergy between the application data transfers and the che
 ckpointing transfers can lead to a reasonable additional network load, mea
 sured to be lower than +10% on a dense linear algebra example.\n\nTag: Ext
 reme Scale Computing, Fault Tolerance, Reliability and Resiliency\n\nRegis
 tration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR