BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20190719T085745Z
LOCATION:HG E 1.2
DTSTART;TZID=Europe/Stockholm:20190614T140000
DTEND;TZID=Europe/Stockholm:20190614T143000
UID:submissions.pasc-conference.org_PASC19_sess144_msa218@linklings.com
SUMMARY:Nonlinear Multi-Physics 3-D Solver: from CUDA C + MPI to Julia
DESCRIPTION:Minisymposium\nComputer Science and Applied Mathematics, Physi
cs, Solid Earth Dynamics\n\nNonlinear Multi-Physics 3-D Solver: from CUDA
C + MPI to Julia\n\nRĂ¤ss, Omlin, Podladchikov\n\nWe present first results
from a Julia port of a massively parallel Multi-GPU solver that is origina
lly written in CUDA C using MPI. It resolves spontaneous nonlinear multi-p
hysics flow localisation in 3-D. The solver's parallel, matrix-free design
enables a short time to solution and its applicability spans over a wide
range of coupled and nonlinear systems of partial differential equations i
n 3-D. The used stencil-based iterative method is optimally suited for bot
h shared and distributed memory parallelism. The original Multi-GPU solver
achieved a high performance and a nearly ideal parallel efficiency on up
to 5120 NVIDIA Tesla P100 GPUs on the hybrid Cray XC-50 "Piz Daint" superc
omputer at the Swiss National Supercomputing Centre, CSCS. We share the fi
rst performance and scaling results obtained with the Julia port and evalu
ate the correctness of simulation results. We report challenges and discus
s future paths.
END:VEVENT
END:VCALENDAR