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:20190719T085745Z
LOCATION:HG D 1.1
DTSTART;TZID=Europe/Stockholm:20190614T110000
DTEND;TZID=Europe/Stockholm:20190614T113000
UID:submissions.pasc-conference.org_PASC19_sess119_msa175@linklings.com
SUMMARY:Improving Ocean Model Computational Performance by using Mixed-Pre
 cision Approaches
DESCRIPTION:Minisymposium\nClimate and Weather\n\nImproving Ocean Model Co
 mputational Performance by using Mixed-Precision Approaches\n\nTinto, Acos
 ta, Castrillo, Serradell, Doblas-Reyes\n\nMixed-precision approaches can p
 rovide substantial speed-ups for both computing- and memory-bound codes re
 quiring little effort. Most scientific codes have overengineered the numer
 ical precision leading to a situation where models are using more resource
 s than required without having a clue about where these resources are unne
 cessary and where are really needed. Consequently, there is the possibilit
 y to obtain performance benefits from using a more appropriate choice of p
 recision and the only thing that is needed is a method to determine which 
 real variables can be represented with fewer bits without affecting the ac
 curacy of the results. We developed a novel method that can enable modern 
 and legacy codes to benefit from a reduction of precision without sacrific
 ing accuracy. It consists in a simple idea: if we can measure how reducing
  the precision of a group of variables affects the outputs, we can evaluat
 e the level of precision this group of variables need. The method has been
  put to proof using two state-of-the-art ocean models, NEMO and ROMS, with
  very promising results. A comparative performance analysis between the mi
 xed- and double-precision versions of the code offers insight about the be
 nefits of mixed-precision.
END:VEVENT
END:VCALENDAR