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:20230124T170804Z
LOCATION:C1-2-3
DTSTART;TZID=America/Chicago:20221116T083000
DTEND;TZID=America/Chicago:20221116T170000
UID:submissions.supercomputing.org_SC22_sess225_spostg116@linklings.com
SUMMARY:IOPathTune: Adaptive Online Parameter Tuning for Parallel File Sys
 tem I/O Path
DESCRIPTION:ACM Student Research Competition: Graduate Poster, ACM Student
  Research Competition: Undergraduate Poster, Posters\n\nIOPathTune: Adapti
 ve Online Parameter Tuning for Parallel File System I/O Path\n\nRashid\n\n
 Parallel file systems like Lustre contain complicated I/O paths from clien
 ts to storage servers. An efficient I/O path requires proper settings of m
 ultiple parameters as the default settings often fail to deliver optimal p
 erformance, especially for diverse workloads in the HPC environment. Exist
 ing tuning strategies are limited in being adaptive, timely, and flexible.
  We propose IOPathTune, which adaptively tunes PFS I/O Path online from th
 e client side without characterizing the workloads, doing expensive profil
 ing, and communicating with other machines. We leveraged CloudLab to condu
 ct the evaluations with 20 different Filebench workloads under three diffe
 rent test conditions: single-client standalone tests, dynamic workload cha
 nge, and multi-client executions. We observed either on par or better perf
 ormance than the default configuration across all workloads. Some of the m
 ost considerable improvement includes 231%, 113%, 96%, 43%.\n\nRegistratio
 n Category: Tech Program Reg Pass, Exhibits Reg Pass
END:VEVENT
END:VCALENDAR
