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:D165
DTSTART;TZID=America/Chicago:20181112T163000
DTEND;TZID=America/Chicago:20181112T170000
UID:submissions.supercomputing.org_SC18_sess161_ws_pmbsf120@linklings.com
SUMMARY:Evaluating the Impact of Spiking Neural Network Traffic on Extreme
 -Scale Hybrid Systems
DESCRIPTION:Workshop\nBenchmarks, Parallel Programming Languages, Librarie
 s, and Models, Performance, Simulation, Workshop Reg Pass\n\nEvaluating th
 e Impact of Spiking Neural Network Traffic on Extreme-Scale Hybrid Systems
 \n\nWolfe, Plagge, Mubarak, Carothers, Ross\n\nAs we approach the limits o
 f Moore's law, there is increasing interest in non-Von Neuman architecture
 s such as neuromorphic computing to take advantage of improved compute and
  low power capabilities. Spiking neural network (SNN) applications have so
  far shown very promising results running on a number of processors, motiv
 ating the desire to scale to even larger systems having hundreds and even 
 thousands of neuromorphic processors. Since these architectures currently 
 do not exist in large configurations, we use simulation to scale real neur
 omorphic applications running on a single neuromorphic chip, to thousands 
 of chips in an HPC class system. Furthermore, we use a novel simulation wo
 rkflow to perform a full scale systems analysis of network performance and
  the interaction of neuromorphic workloads with traditional CPU workloads 
 in a hybrid supercomputer environment. On average, we find Slim Fly, Fat-T
 ree, Dragonfly-1D, and Dragonfly-2D are 45%, 46%, 76%, and 83% respectivel
 y faster than the worst case performing topology for both convolutional an
 d Hopfield NN workloads running alongside CPU workloads. Running in parall
 el with CPU workloads translates to an average slowdown of 21% for a Hopfi
 eld type workload and 184% for convolutional NN workloads across all HPC n
 etwork topologies.
URL:https://sc18.supercomputing.org/presentation/?id=ws_pmbsf120&sess=sess
 161
END:VEVENT
END:VCALENDAR