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:20260605T154540Z
LOCATION:Plenary Room (Bldg. 6 - 001)
DTSTART;TZID=Europe/Stockholm:20260629T194600
DTEND;TZID=Europe/Stockholm:20260629T194700
UID:submissions.pasc-conference.org_PASC26_sess124_pos134@linklings.com
SUMMARY:P25 - Maintainable, Sustainable, and Generalizable Data Layouts an
 d Vectorization for Rigid-Body Molecular Dynamics
DESCRIPTION:Samuel James Newcome, Luis Gall, David Martin, Markus Mühlhäuß
 er, and Hans-Joachim Bungartz (Technical University of Munich)\n\nls1-MarD
 yn (ls1) is an MD simulator designed for large-scale simulations of multi-
 site molecules and has been successfully used in a variety of scientific s
 tudies. It represents molecules as rigid bodies composed of multiple inter
 action sites that each exert forces on their neighbours, which are determi
 ned by those molecules with centre-of-masses within some cutoff distance. 
 \n\nPast works have integrated the modern algorithm-selection particle sim
 ulation library, AutoPas, into ls1, replacing its older particle container
  and force calculation mechanisms. However, the limitations of AutoPas con
 fine this integration to only single-site molecules. Extending these conce
 pts to multi-site molecular representations introduces additional challeng
 es, such as enabling efficient vectorization. At the same time, this highl
 ights a broader issue, namely how to maintain high performance while ensur
 ing long-term code sustainability across different hardware architectures.
 \n\nTo address this, we focus on improving the maintainability and portabi
 lity of vectorized force calculations. Instead of relying on manually spec
 ialized implementations for each instruction set, we introduce Google High
 way as a portable SIMD abstraction layer. This approach significantly redu
 ces code complexity while retaining competitive performance. We present in
 itial results and discuss practical limitations.\n\nSession Chair: Mirosla
 va Nedyalkova (University of Fribourg)\n\n
END:VEVENT
END:VCALENDAR