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:20181221T160728Z
LOCATION:D166
DTSTART;TZID=America/Chicago:20181112T143000
DTEND;TZID=America/Chicago:20181112T150000
UID:submissions.supercomputing.org_SC18_sess173_ws_espm109@linklings.com
SUMMARY:A Unified Runtime for PGAS and Event-Driven Programming
DESCRIPTION:Workshop\nAccelerators, Exascale, Parallel Programming Languag
 es, Libraries, and Models, Workshop Reg Pass\n\nA Unified Runtime for PGAS
  and Event-Driven Programming\n\nPaul, Chen, Hayashi, Grossman, Sarkar\n\n
 A well-recognized characteristic of extreme scale systems is that their co
 mputation bandwidths far exceed their communication bandwidths.  PGAS runt
 imes have proven to be effective in enabling efficient use of communicatio
 n bandwidth, due to their efficient support for short nonblocking one-side
 d messages.  However, they were not designed for exploiting the massive le
 vels of intra-node parallelism found in extreme scale systems.\n\nIn this 
 paper, we explore the premise that event-driven intra-node runtimes could 
 be promising candidates for integration with PGAS runtimes, due to their a
 bility to overlap computation with long-latency operations.  For this, we 
 use OpenSHMEM as an exemplar of PGAS runtimes, and Node.js as an exemplar 
 of event-driven runtimes.  While Node.js may seem an unusual choice for hi
 gh-performance computing, its prominent role as an event-driven runtime fo
 r server-side Javascript is a good match for what we need for optimizing t
 he use of communication bandwidth.  Availability of excess computation ban
 dwidth in extreme scale systems can help mitigate the local computation ov
 erheads of Javascript.  Further, since the Node.js environment is single-t
 hreaded, we get an automatic guarantee that no data races will occur on Ja
 vascript objects, a guarantee that we cannot get with C++.\n\nOur integrat
 ion of OpenSHMEM and Node.js makes it possible to expose nonblocking PGAS 
 operations as JavaScript constructs that can seamlessly be used with the J
 avaScript asynchronous mechanisms, such as those based on promises.  We be
 lieve that the exploration and preliminary results in this paper offer a n
 ew direction for future research on building high productivity runtimes fo
 r extreme scale systems.
URL:https://sc18.supercomputing.org/presentation/?id=ws_espm109&sess=sess1
 73
END:VEVENT
END:VCALENDAR