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:20260421T090513Z
LOCATION:Plenary Room (Bldg. 6 - 001)
DTSTART;TZID=Europe/Stockholm:20260629T193800
DTEND;TZID=Europe/Stockholm:20260629T193900
UID:submissions.pasc-conference.org_PASC26_sess124_pos108@linklings.com
SUMMARY:GPU-Accelerated Methods for Numerically Stable Resampling in Fluid
 -Structure Interaction
DESCRIPTION:Simone Riva (Università della Svizzera italiana) and Patrick Z
 ulian (UniDistance Suisse, Università della Svizzera italiana)\n\nFluid-st
 ructure interaction simulations require accurate transfer of scalar fields
  between overlapping meshes with different topologies. We address the prob
 lem of transferring fields from unstructured tetrahedral to structured hex
 ahedral meshes.\n\nThis problem is challenging because direct quadrature m
 ethods suffer from numerical instability due to hexahedral grid undersampl
 ing, while more sophisticated solutions are difficult to parallelize on GP
 Us efficiently.\nWe developed and compared three GPU-optimized approaches,
  tested on the Alps GH200 at CSCS, analyzing their trade-offs in stability
 , accuracy, and performance.\n\nThe first approach uses iterative refineme
 nt of tetrahedral quadrature rules, achieving high accuracy but with limit
 ed GPU parallelism. The second employs geometric adaptivity through explic
 it refinement rules instead of recursion, demonstrating high GPU throughpu
 t. The third samples use hexahedral quadrature rules, which is ideal for G
 PU parallelization but creates topological conflicts when multiple tetrahe
 dra contribute to single quadrature nodes. We resolve this using an adapte
 d Cell-List data structure that efficiently handles conflicts while mainta
 ining near-optimal GPU performance.\n\nBy combining geometric refinement w
 ith GPU-optimized spatial indexing, we achieve both numerical stability an
 d parallel efficiency. Our methods enable robust, high-performance mesh tr
 ansfer for fluid-structure interaction and other multiphysics frameworks c
 oupling different mesh topologies.\n\n
END:VEVENT
END:VCALENDAR