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:20260522T150115Z
LOCATION:C2/3/4 Ballroom
DTSTART;TZID=America/Chicago:20181113T083000
DTEND;TZID=America/Chicago:20181113T170000
UID:submissions.supercomputing.org_SC18_sess322_post126@linklings.com
SUMMARY:An Efficient SIMD Implementation of Pseudo-Verlet Lists for Neighb
 or Interactions in Particle-Based Codes
DESCRIPTION:James Willis (Durham University, Institute for Computational C
 osmology); Matthieu Schaller (Leiden Observatory); and Pedro Gonnet (Googl
 e LLC)\n\nIn particle-based simulations, neighbour finding (i.e. finding p
 airs of particles to interact within a given range) is the most time consu
 ming part of the computation. One of the best such algorithms, which can b
 e used for both Molecular Dynamics (MD) and Smoothed Particle Hydrodynamic
 s (SPH) simulations is the pseudo-Verlet list algorithm. The algorithm imp
 roves the neighbour finding by reducing the number of spurious pair-wise d
 istance calculations. This algorithm, however, does not vectorize triviall
 y, and hence makes it difficult to exploit SIMD-parallel architectures. On
  this poster, we present several novel modifications as well as a vectoriz
 ation strategy for the algorithm which lead to overall speed-ups over the 
 scalar version of the algorithm of 2.21x for the AVX instruction set (SIMD
  width of 8), 2.41x for AVX2, and 3.65x for AVX-512 (SIMD width of 16).\n\
 nRegistration Category: Tech Program Reg Pass, Exhibits Reg Pass\n\n
END:VEVENT
END:VCALENDAR