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:20190731T213612Z
LOCATION:Crystal C
DTSTART;TZID=America/Chicago:20190730T120000
DTEND;TZID=America/Chicago:20190730T123000
UID:pearc_PEARC19_sess202_stu_pap115@linklings.com
SUMMARY:Advanced Research Computing Containers for AI Workloads
DESCRIPTION:Advanced Research Computing Containers for AI Workloads\n\nSma
 thers, Newlin, DeYoung\n\nContainerization has taken the software world by
  storm. Deployment\ncomplications, like requiring elevated (i.e. “root”) p
 ermissions\nto run, have slowed the adoption of containers in shared\nadva
 nced research computing (ARC) environments. Singularity is a\ncontaineriza
 tion approach that is designed for ARC in shared high\nperformance computi
 ng (HPC) clusters. With the creation of the\nSingularity, there is finally
  a viable scientific container solution.\nHowever very few papers have loo
 ked at the performance tradeoffs\nof deploying applications using a contai
 ner based model. The\nauthors are not aware of any published studies evalu
 ating the tradeoffs\nof the deployment models with complex Artificial Inte
 lligence\n(AI) workloads. Without detailed evaluations of the performance\
 ntrade-offs scientists and engineers are unable to make an informed\ndecis
 ion on deployment model for time sensitive training or low\npower inferenc
 e. Furthering previous research in this area and\nusing emerging community
  developed benchmarks, we examine\nperformance trade-offs of running AI wo
 rkloads in a containerized\nSingularity environment.\n\nPaper\nStudent Tec
 hnical Paper, Software, Applications
URL:https://pearc19.conference-program.com/presentation/?id=stu_pap115&ses
 s=sess202
END:VEVENT
END:VCALENDAR

