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:20260605T154541Z
LOCATION:Bldg. 6 - Room 104
DTSTART;TZID=Europe/Stockholm:20260701T140000
DTEND;TZID=Europe/Stockholm:20260701T160000
UID:submissions.pasc-conference.org_PASC26_sess163@linklings.com
SUMMARY:MS5F - Single and Reduced Precision for Weather and Climate Models
DESCRIPTION:Organizer(s): Balthasar Reuter (ECMWF), and Claudia Frauen (DK
 RZ)\n\nHigh-resolution weather and climate models are rapidly approaching 
 the long-standing goal of simulating one year per day at 1 km global resol
 ution, thanks to GPU-accelerated pre-Exascale and Exascale HPC systems. Ac
 hieving this performance sustainably, however, requires more than traditio
 nal code optimization - it demands HPC co-design, where scientific develop
 ers, numerical analysts, and computational experts collaborate to rethink 
 algorithms for modern hardware. One promising strategy is the use of singl
 e- or reduced-precision arithmetic. Lower-precision computations improve m
 emory bandwidth, cache efficiency, and energy usage, and enable exploitati
 on of specialized GPU units optimized for fast, low-precision operations. 
 At the same time, reducing precision raises important questions about scie
 ntific trust, numerical stability, and reproducibility, particularly in ch
 aotic, multi-scale Earth system models. This minisymposium will bring toge
 ther experts from climate science, numerical modeling, and HPC to share ex
 periences, challenges, and best practices in adopting lower-precision comp
 utations. Topics include algorithmic reformulation, precision-aware softwa
 re design, verification and validation strategies, and precision analysis 
 tools. While focused on weather and climate models, the lessons learned ar
 e broadly applicable across computational science, offering guidance for b
 uilding performant, trustworthy, and future-ready models on next-generatio
 n HPC platforms.\n\nDoes Halving the Precision Yield Twice the Speed?\n\nM
 odern GPU architectures offer substantially higher throughput for single-p
 recision arithmetic compared to double precision, making reduced-precision
  computation an attractive strategy for accelerating weather and climate m
 odels. In theory, halving the precision could yield up to twice the perfor
 man...\n\n\nDmitry Alexeev (NVIDIA Inc.)\n---------------------\nReducing 
 and Validating Precision in the ICON model\n\nReducing model precision can
  significantly reduce the memory footprint of a model, and increase utiliz
 ation on future AI-based hardware. However, identifying and validating sin
 gle precision model components without compromising on the model accuracy 
 of scientific experiments remains a key challenge....\n\n\nDylan Kierans (
 DKRZ)\n---------------------\nPerformance, Precision and Accuracy: Mixed-P
 recision Modelling for Weather and Climate Applications at the MetOffice\n
 \nNumerical calculations using floating point numbers suffer from accumula
 ted round-off error and moreover, they are not associative. This is proble
 matic for  weather and climate applications as they rely on global sums in
  PDE solvers and the numerical answer can then depend on the parallel deco
 mposit...\n\n\nChristopher Michael Maynard (Met Office)\n-----------------
 ----\nSingle Precision Trajectories in 4D‑Var Data Assimilation\n\nFour‑di
 mensional variational (4D‑Var) data assimilation is underpinning the gener
 ation of high‑quality initial conditions required for ECMWF’s weather fore
 casts and Destination Earth's (DestinE) Digital Twins. A major cost driver
  in this system is the nonlinear (NL) model tra...\n\n\nJorge Bandeiras, P
 atrick Gillies, Balthasar Reuter, and Ioan Hadade (ECMWF)\n\nDomain: Clima
 te, Weather, and Earth Sciences\n\nSession Chairs: Balthasar Reuter (ECMWF
 ) and Claudia Frauen (DKRZ)
END:VEVENT
END:VCALENDAR