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:20181221T160730Z
LOCATION:C140/142
DTSTART;TZID=America/Chicago:20181114T113000
DTEND;TZID=America/Chicago:20181114T120000
UID:submissions.supercomputing.org_SC18_sess188_pap537@linklings.com
SUMMARY:Characterization of MPI Usage on a Production Supercomputer
DESCRIPTION:Paper\nArchitectures, MPI, Networks, Performance, Programming 
 Systems, State of the Practice, Tech Program Reg Pass\n\nCharacterization 
 of MPI Usage on a Production Supercomputer\n\nChunduri, Parker, Balaji, Ha
 rms, Kumaran\n\nMPI is the most prominent programming model used in scient
 ific computing today.  Despite it's importance, however, how scientific ap
 plications use it in production is not very well understood due to the lac
 k of low overhead profiling tools.   We used a lightweight profiling tool,
  called autoperf, to log the MPI usage characteristics of production appli
 cations on a large supercomputing system (Mira) and its corresponding deve
 lopment system (Cetus).  Autoperf limits the amount of information that it
  records in order to keep the overhead to a minimum while still storing en
 ough data to derive useful insights.  MPI usage statistics have been colle
 cted for over 100K jobs that were run within a 2-year period and are analy
 zed.  The analysis of this data is intended as a mechanism to provide usef
 ul insights for MPI developers and network hardware developers for their n
 ext generation of improvements, and for supercomputing center operators fo
 r their next system procurements.
URL:https://sc18.supercomputing.org/presentation/?id=pap537&sess=sess188
END:VEVENT
END:VCALENDAR