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:20230831T095746Z
LOCATION:Sertig
DTSTART;TZID=Europe/Stockholm:20230628T110000
DTEND;TZID=Europe/Stockholm:20230628T113000
UID:submissions.pasc-conference.org_PASC23_sess165_msa185@linklings.com
SUMMARY:Implementing GPU Support for Unstructured Ocean Model FESOM2 and C
 oupled IFS/FESOM2 Model
DESCRIPTION:Minisymposium\n\nWilton Jaciel Loch and Enrico Degregori (Germ
 an Climate Computing Centre) and Dmitry Sidorenko, Nikolay Koldunov, Patri
 ck Scholz, Vera Fofonova, Sergey Danilov, and Thomas Jung (AWI)\n\nClimate
  configuration being used employs a sequential coupling approach where the
  FESOM and IFS models are run as a single executable and share the same in
 frastructure consisting of MPI+OpenMP+OpenACC. IFS model is already benefi
 ting from GPU acceleration for performing the spectral transform, while th
 e FESOM model required extension to support GPU acceleration in order to f
 ully leverage the hardware capabilities in the climate setup. Hence, the F
 ESOM 2.1 tracer and sea ice components were successfully ported to the DKR
 Z Levante machine using the OpenACC approach for offloading, in close coll
 aboration with the DKRZ team. Analysis of the potential benefits of using 
 GPUs revealed that optimization of the model code is necessary for effecti
 ve use. The optimized code was evaluated for performance improvements on G
 PUs and degradation on CPUs. Results indicate a significant speedup of up 
 to a factor of 5 on the GPU cluster, with minimal degradation on CPUs. Whi
 le not all of the FESOM 2.1 code has been ported, initial findings suggest
  that using OpenACC to port and optimize the model for GPU is a viable app
 roach. Our study presents the results of a comparison between CPU and GPU 
 performance improvements.\n\nDomain: Climate, Weather and Earth Sciences\n
 \nSession Chair: Xavier Lapillonne (MeteoSwiss)
END:VEVENT
END:VCALENDAR