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:507
DTSTART;TZID=America/Denver:20191117T121000
DTEND;TZID=America/Denver:20191117T123000
UID:submissions.supercomputing.org_SC19_sess108_ws_pawatm114@linklings.com
SUMMARY:Scalable Machine Learning with OpenSHMEM
DESCRIPTION:Workshop\n\nScalable Machine Learning with OpenSHMEM\n\nTaylor
 , Dinan, Rahman, Ozog\n\nDeep convolutional neural networks (DNNs) have ha
 d a significant, and lasting impact across the computing industry. Trainin
 g these large neural networks is computationally intensive and is often pa
 rallelized to shorten training times that could otherwise range from days 
 to weeks. The Message Passing Interface (MPI) communication model has been
  commonly used to facilitate the data exchange and synchronization require
 d for parallel DNN training. We observe that OpenSHMEM supports many of th
 e same communication operations as MPI — in particular, the all-reduce ope
 ration needed to support data parallelism — and that OpenSHMEM may further
  provide a unique solution to fine-grain model parallel computation. In th
 is work, we present an initial evaluation of OpenSHMEM’s suitability for u
 se in DNN training and compare its performance with MPI. Results indicate 
 that OpenSHMEM data-parallel performance is comparable with MPI. The usage
  of OpenSHMEM to support model parallelism will be explored in our future 
 work.\n\nTag: Workshop Reg Pass, Machine Learning, MPI, Parallel Applicati
 on Frameworks, Parallel Programming Languages, Libraries, and Models, Scal
 able Computing\n\nRegistration Category: Workshop Reg Pass, Machine Learni
 ng, MPI, Parallel Application Frameworks, Parallel Programming Languages, 
 Libraries, and Models, Scalable Computing
URL:https://sc19.supercomputing.org/presentation/?id=ws_pawatm114&sess=ses
 s108
END:VEVENT
END:VCALENDAR