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:20211207T055355Z
LOCATION:Second Floor Atrium
DTSTART;TZID=America/Chicago:20211117T083000
DTEND;TZID=America/Chicago:20211117T170000
UID:submissions.supercomputing.org_SC21_sess279_rpost111@linklings.com
SUMMARY:Support in OpenMP for Multi-GPU Parallelism
DESCRIPTION:Posters, Research Posters\n\nSupport in OpenMP for Multi-GPU P
 arallelism\n\nTorres, Kale, Malik, Scogland, Ferrer...\n\nNodes of emergin
 g supercomputers have multiple GPUs, i.e., a multi-GPU, on them. Applicati
 ons are often parallelized across the GPUs of a multi-GPU using MPI, but a
  more performant and portable solution for parallelizing across the GPUs i
 s needed. OpenMP, which is used to parallelize computation within a multi-
 core or a GPU, could facilitate parallelization of computation across the 
 GPUs in a performant and portable way through, e.g., low memory requiremen
 ts compared to MPI and directive-based parallelization. In this work, we p
 resent a solution that provides support in OpenMP for parallelizing an app
 lication across GPUs of a multi-GPU through language extensions and compil
 er optimizations developed in LLVM's OpenMP implementation. Preliminary ex
 perimentation of our solution using the Stream benchmark on a cluster’s no
 de having four GPUs suggests that our approach can be a performant, portab
 le and easy-to-use solution for application programmers to harness the com
 putational power of the GPUs of a node.\n\nRegistration Category: Tech Pro
 gram Reg Pass, Exhibit Hall Only
END:VEVENT
END:VCALENDAR