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:20210402T160045Z
LOCATION:Track 5
DTSTART;TZID=America/New_York:20201117T100000
DTEND;TZID=America/New_York:20201117T103000
UID:submissions.supercomputing.org_SC20_sess169_pap222@linklings.com
SUMMARY:Improving All-to-Many Personalized Communication in Two-Phase I/O
DESCRIPTION:Paper\n\nImproving All-to-Many Personalized Communication in T
 wo-Phase I/O\n\nKang, Ross, Latham, Lee, Agrawal...\n\nAs modern parallel 
 computers enter the exascale era, the communication cost for redistributin
 g requests becomes a significant bottleneck in MPI-IO routines. The commun
 ication kernel for request redistribution, which has an all-to-many person
 alized communication pattern for application programs with a large number 
 of noncontiguous requests, plays an essential role in the overall performa
 nce. This paper explores the available communication kernels for two-phase
  I/O communication. We generalize the spread-out algorithm to adapt to the
  all-to-many communication pattern of two-phase I/O by reducing the commun
 ication straggler effect. Communication throttling methods that reduce com
 munication contention for asynchronous MPI implementation are adopted to i
 mprove communication performance further. Experimental results are present
 ed using different communication kernels running on Cray XC40 Cori and IBM
  AC922 Summit supercomputers with different I/O patterns. Our study shows 
 that adjusting communication kernel algorithms for different I/O patterns 
 can improve the end-to-end performance up to 2x compared with default MPI-
 IO implementations.\n\nTag: Algorithms, File Systems and I/O, Storage\n\nR
 egistration Category: Tech Program Reg Pass
END:VEVENT
END:VCALENDAR

