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:20250822T115810Z
LOCATION:Room 5.2D02
DTSTART;TZID=Europe/Stockholm:20250617T123000
DTEND;TZID=Europe/Stockholm:20250617T130000
UID:submissions.pasc-conference.org_PASC25_sess134_msa241@linklings.com
SUMMARY:Sustainable, Trustworthy Coupled HPC+AI for Molecular Simulation a
 nd Materials Design: Energy Consumption, Correctness, and Efficient Traini
 ng on Leadership Platforms
DESCRIPTION:Ada Sedova (Oak Ridge National Laboratory)\n\nThe promise of a
 ccelerating and advancing molecular simulation and materials design effort
 s with coupled HPC and deep learning (DL) workflows has motivated an explo
 sion in a variety of approaches. In particular, leadership computing facil
 ities have supported a diverse set of large-scale efforts in this area. Bu
 t with the increasing size of models and advanced active learning workflow
 s for training, which are arising in response to the need for improvements
  in accuracy and reliability of model predictions, concerns emerge about e
 xcessive energy consumption and the sustainability of HPC+AI simulation ef
 forts for science. In this talk, I will describe experiences developing, d
 eploying and assessing the results of leadership-scale HPC+DL efforts in m
 odeling for molecular and materials sciences, from biosciences to advanced
  materials and nuclear energy, and using several different national leader
 ship supercomputing resources. Successes, pain points, and lessons learned
  will be described, as well as tools being developed to help monitor robus
 tness, correctness and reproducibility as well as power and energy metrics
  across software stack layers and parallel resources.\n\nDomain: Chemistry
  and Materials, Climate, Weather, and Earth Sciences, Computational Method
 s and Applied Mathematics\n\nSession Chairs: Riccardo Balin (Argonne Natio
 nal Laboratory) and Alessandro Rigazzi (HPE)\n\n
END:VEVENT
END:VCALENDAR