BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20210402T160554Z
LOCATION:Track 4
DTSTART;TZID=America/New_York:20201111T160000
DTEND;TZID=America/New_York:20201111T162000
UID:submissions.supercomputing.org_SC20_sess201_ws_ia102@linklings.com
SUMMARY:Distributed Memory Graph Coloring Algorithms for Multiple GPUs
DESCRIPTION:Workshop\n\nDistributed Memory Graph Coloring Algorithms for M
 ultiple GPUs\n\nBogle, Boman, Devine, Rajamanickam, Slota\n\nGraph colorin
 g is often used in parallelizing scientific computations that run in distr
 ibuted and multi-GPU environments; it identifies sets of independent data 
 that can be updated in parallel. Many algorithms exist for graph coloring 
 on a single GPU or in distributed memory, but hybrid MPI+GPU algorithms ha
 ve been unexplored until this work, to the best of our knowledge. We prese
 nt several MPI+GPU coloring approaches that use implementations of the dis
 tributed coloring algorithms of Gebremedhin et al. and the shared-memory a
 lgorithms of Deveci et al. The on-node parallel coloring uses implementati
 ons in KokkosKernels, which provide parallelization for both multicore CPU
 s and GPUs. We further extend our approaches to solve for distance-2 color
 ing, giving the first known distributed and multi-GPU algorithm for this p
 roblem. In addition, we propose novel methods to reduce communication in d
 istributed graph coloring. Our experiments show that our approaches operat
 e efficiently on inputs too large to fit on a single GPU and scale up to g
 raphs with 76.7 billion edges running on 128 GPUs.\n\nRegistration Categor
 y: Workshop Reg Pass
END:VEVENT
END:VCALENDAR