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:20210916T132447Z
LOCATION:Henry Dunant
DTSTART;TZID=Europe/Stockholm:20210709T113000
DTEND;TZID=Europe/Stockholm:20210709T120000
UID:submissions.pasc-conference.org_PASC21_sess188_pap103@linklings.com
SUMMARY:Performance Optimization and Load-Balancing Modeling for Superpara
 metrization by 3D LES
DESCRIPTION:Paper\n\nPerformance Optimization and Load-Balancing Modeling 
 for Superparametrization by 3D LES\n\nvan den Oord, Chertova, Jansson, Pel
 upessy, Siebesma...\n\nIn order to eliminate climate uncertainty w.r.t clo
 ud and convection parametrizations, superpramaterization (SP) has emerged 
 as one of the possible ways forward. We have implemented (regional) superp
 arametrization of the ECMWF weather model OpenIFS by cloud-resolving, thre
 e-dimensional large-eddy simulations. This setup contains a two-way coupli
 ng between a global meteorological model that resolves large-scale dynamic
 s, with many local instances of the Dutch Atmospheric Large Eddy Simulatio
 n (DALES), resolving cloud and boundary layer physics. The model is curren
 tly prohibitively expensive to run over climate or even seasonal time scal
 es, and a global SP requires the allocation of millions of cores. In this 
 paper, we study the performance and scaling behavior of the LES models and
  the coupling code and present our implemented optimizations. We mimic the
  observed load imbalance with a simple performance model and present strat
 egies to improve hardware utilization in order to assess the feasibility o
 f a world-covering superparametrization. We conclude that (quasi-)dynamica
 l load-balancing can significantly reduce the runtime for such large-scale
  systems with wide variability in LES time-stepping speeds.\n\nDomain: CS 
 and Math, Climate and Weather
END:VEVENT
END:VCALENDAR