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:20190719T085744Z
LOCATION:HG E 1.1
DTSTART;TZID=Europe/Stockholm:20190613T141500
DTEND;TZID=Europe/Stockholm:20190613T144500
UID:submissions.pasc-conference.org_PASC19_sess142_msa268@linklings.com
SUMMARY:From Titan to Summit: Experience in Porting the XGC Gyrokinetic Co
 de to Multi‐GPU Architecture
DESCRIPTION:Minisymposium\nComputer Science and Applied Mathematics, Physi
 cs\n\nFrom Titan to Summit: Experience in Porting the XGC Gyrokinetic Code
  to Multi‐GPU Architecture\n\nEthier, Scheinberg, Abbott, Worley, D’
 Azevedo...\n\nThis presentation will discuss some of the details in portin
 g the gyrokinetic PIC code XGC to the new 200-PF “Summit” comp
 uter at the Oak Ridge Leadership Computing Facility (OLCF). Its multi-GPU 
 node architecture allows for optimizations that were not available on its 
 predecessor, the “Titan” system. The various configurations of
  resource sets give flexibility in running the code with either one GPU pe
 r MPI task, as on Titan, or with full multi-GPU mode to allow for GPU-awar
 e MPI. While improving the CUDA Fortran and OpenACC algorithm currently im
 plemented in XGC, a new porting approach is also being explored by using &
 ldquo;Cabana”, the Kokkos-based library developed as part of the Co-
 Design center for Particle Applications, which is funded by the US DOE Exa
 scale Computing Project.
END:VEVENT
END:VCALENDAR