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:20220812T074334Z
LOCATION:Darwin Room
DTSTART;TZID=Europe/Stockholm:20220627T163000
DTEND;TZID=Europe/Stockholm:20220627T170000
UID:submissions.pasc-conference.org_PASC22_sess130_msa149@linklings.com
SUMMARY:Performance Portability of Lagrange-Stencil-Algorithms using Kokko
 s
DESCRIPTION:Minisymposium\n\nPerformance Portability of Lagrange-Stencil-A
 lgorithms using Kokkos\n\nSchild, Räth, Kormann, Hallatschek\n\nFully kine
 tic plasma simulations using semi-Lagrangian solvers require high memory r
 esources and efficient shared and distributed parallelization schemes. The
  computational most intensiv part of semi-Lagrangian algorithms relies on 
 interpolation algorithms. Using the performance protability framework Kokk
 os a 1d Stencil algorithm including a prefetching scheme is presented. <br
  />The performance of the algorithm on CPU and GPU is evaluated with regar
 d to the roofline model. On CPU up to 90% of the Floating Point Performanc
 e could be achieved depending on the stencil. By switching from CPU to GPU
  a runtime improvment of a factor of 10 could easily be achieved without c
 hanging the implementation. <br />In addition the creation of the template
 d interpolation class is shown using a Builder Pattern to enhance the flex
 ibility of the implemented algorithm with regards to different stencils an
 d physical models used in the interpolation.\n\nDomain: Computer Science a
 nd Applied Mathematics, Engineering, Physics
END:VEVENT
END:VCALENDAR