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:20230124T171525Z
LOCATION:C147-154
DTSTART;TZID=America/Chicago:20221113T110000
DTEND;TZID=America/Chicago:20221113T113000
UID:submissions.supercomputing.org_SC22_sess427_ws_exampi101@linklings.com
SUMMARY:Compiler-Enabled Optimization of Persistent MPI Operations
DESCRIPTION:Workshop\n\nCompiler-Enabled Optimization of Persistent MPI Op
 erations\n\nJammer, Bishcof\n\nMPI also includes persistent operations, wh
 ich specify recurring communication patterns.  The idea is that the usage 
 of those operations can result in a performance benefit compared to the st
 andard non-blocking communication.  But in current MPI implementations thi
 s performance benefit is not really observable.  We determine the message 
 envelope matching as one of the causes of overhead.  As persistent MPI req
 uests can be used multiple times, the compiler can, in some cases, prove t
 hat message matching is only needed for the first occurrence and can be en
 tirely skipped for subsequent usages.\n\nWe present the required compiler 
 analysis and an implementation of a communication scheme that skips the me
 ssage envelope matching.  This allows us to substantially reduce the commu
 nication overhead that cannot be overlapped with computation.  Using the I
 ntel IMB-ASYNC Benchmark, we can see a communication overhead reduction of
  up to 95% for larger message sizes.\n\nSession Format: Recorded\n\nRegist
 ration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR