BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20200129T163559Z
LOCATION:605
DTSTART;TZID=America/Denver:20191117T143000
DTEND;TZID=America/Denver:20191117T150000
UID:submissions.supercomputing.org_SC19_sess109_ws_exampi104@linklings.com
SUMMARY:Using MPI-3 RMA for Active Messages
DESCRIPTION:Workshop\n\nUsing MPI-3 RMA for Active Messages\n\nSchuchart, 
 Bouteiller, Bosilca\n\nDistributed asynchronous programming systems requir
 e a scalable, low-latency way to exchange information on the state of unit
 s of execution (tasks) across process boundaries. These interactions can b
 e modeled as active messages, i.e., a message that, upon reception, trigge
 rs an action on the receiving side. Such actions, generally of short durat
 ion, include the posting of the reception for a task input data or the tra
 nsition of a task into the runnable state. Traditionally, such messages ha
 ve been implemented using MPI two-sided communication primitives. However,
  processing reception queues for numerous small messages may incur signifi
 cant overheads in the progress of the MPI library, notably when handling u
 nexpected messages and iterating over a number of pre-posted requests. In 
 this work, we investigate a different approach for implementing an active 
 message queue using MPI-3 RMA primitives and compare the resulting perform
 ance in a task runtime in terms of scalable performance and latency agains
 t two-sided implementations. We discuss the latency of basic RMA operation
 s and the resulting performance in terms of latency and scalable throughpu
 t of the RMA-based queues. Using a benchmark application, we show that usi
 ng RMA-based queues may provide notable benefits compared to using MPI two
 -sided communication for inter-scheduler\ncommunication.\n\nTag: Workshop 
 Reg Pass, Exascale, MPI, Networks, Parallel Programming Languages, Librari
 es, and Models\n\nRegistration Category: Workshop Reg Pass, Exascale, MPI,
  Networks, Parallel Programming Languages, Libraries, and Models
URL:https://sc19.supercomputing.org/presentation/?id=ws_exampi104&sess=ses
 s109
END:VEVENT
END:VCALENDAR