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:20230831T095747Z
LOCATION:Wisshorn
DTSTART;TZID=Europe/Stockholm:20230628T143000
DTEND;TZID=Europe/Stockholm:20230628T150000
UID:submissions.pasc-conference.org_PASC23_sess166_msa272@linklings.com
SUMMARY:PhyDLL - Physics Deep Learning CoupLer: An Open-Source High-Perfor
 mance Coupling Library
DESCRIPTION:Minisymposium\n\nAnass Serhani, Corentin Lapeyre, and Gabriel 
 Staffelbach (CERFACS)\n\nThere has been recently a significant shift towar
 ds using data-driven approaches, particularly deep learning (DL) in comput
 ational fluid dynamics (CFD). While these techniques have shown great prom
 ise in improving the accuracy of fluid simulations, there is a growing nee
 d to consider the high-performance efficiency and robustness when solvers 
 are querying online DL inference. To address these concerns, the open-sour
 ce coupling library PhyDLL (Physical Deep Learning coupLer) has been devel
 oped. It allows high-performance data transfer and processing between mass
 ively-parallel fluid solvers and distributed DL inferences. This library i
 ncludes different coupling schemes that fit the context encompassing the C
 FD solver and DL engine. PhyDLL provides therefore a Fortran interface, th
 at cater to a large portion of CFD solvers, as well as a Python one which 
 aligns well with prime DL libraries such as Tensorflow and Pytorch. Toward
  exascale, PhyDLL is well designed to take advantage of computing capabili
 ties of hybrid architecture (CPU-GPU) of modern clusters. A C/C++ interfac
 e will also be available for better portability and to reach an even wider
  range of users. Finally, significant physical and performance results hav
 e been achieved using PhyDLL to couple the CFD solver AVBP to graph and co
 nvolutional neural network for combustion and aerodynamic use-cases.\n\nDo
 main: Physics\n\nSession Chair: Marta Garcia-Gasulla (Barcelona Supercompu
 ting Center)
END:VEVENT
END:VCALENDAR