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:20260522T150117Z
LOCATION:A2 Ballroom
DTSTART;TZID=America/Chicago:20181114T163000
DTEND;TZID=America/Chicago:20181114T170000
UID:submissions.supercomputing.org_SC18_sess466_gb105@linklings.com
SUMMARY:Exascale Deep Learning for Climate Analytics
DESCRIPTION:Thorsten Kurth (Lawrence Berkeley National Laboratory), Sean T
 reichler and Joshua Romero (Nvidia Corporation), Mayur Mudigonda (Lawrence
  Berkeley National Laboratory), Nathan Luehr and Everett Phillips (Nvidia 
 Corporation), Ankur Mahesh (Lawrence Berkeley National Laboratory), Michae
 l Matheson (Oak Ridge National Laboratory), Jack Deslippe (Lawrence Berkel
 ey National Laboratory), Massimiliano Fatica (Nvidia Corporation), Mr Prab
 hat (Lawrence Berkeley National Laboratory), and Michael Houston (Nvidia C
 orporation)\n\nWe extract pixel-level masks of extreme weather patterns us
 ing variants of Tiramisu and DeepLabv3+ neural networks. We describe impro
 vements to the software frameworks, input pipeline, and the network traini
 ng algorithms necessary to efficiently scale deep learning on the Piz Dain
 t and Summit systems. The Tiramisu network scales to 5300 P100 GPUs with a
  sustained throughput of 21.0 PF/s and parallel efficiency of 79.0%. DeepL
 abv3+ scales up to 27360 V100 GPUs with a sustained throughput of 325.8 PF
 /s and a parallel efficiency of 90.7% in single precision. By taking advan
 tage of the FP16 Tensor Cores, a half-precision version of the DeepLabv3+ 
 network achieves a peak and sustained throughput of 1.13 EF/s and 999.0 PF
 /s respectively.\n\n
END:VEVENT
END:VCALENDAR