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:20190719T085744Z
LOCATION:HG D 1.1
DTSTART;TZID=Europe/Stockholm:20190612T163000
DTEND;TZID=Europe/Stockholm:20190612T170000
UID:submissions.pasc-conference.org_PASC19_sess147_msa286@linklings.com
SUMMARY:On using a DSL Approach Performance Portability of the LFRic Weath
 er and Climate Model
DESCRIPTION:Minisymposium\nComputer Science and Applied Mathematics, Clima
 te and Weather\n\nOn using a DSL Approach Performance Portability of the L
 FRic Weather and Climate Model\n\nMaynard\n\nLFRic is the next generation 
 weather and climate model being developed to replace the Met Office's Unif
 ied Model. It hosts a new dynamical core, Gung Ho, which employs a mixed f
 inite element on an unstructured mesh discretisation to allow the model to
  scale to a very large degree of parallelism. The model is written in Obje
 ct Oriented Fortran 2003 with a strong separation of concerns between the 
 science code and the parallel/performance code. Combined with metadata abo
 ut the science code, this allows the parallel code to be automatically gen
 erated using a python code called PSyclone. The APIs between the Fortran c
 ode and the code generator form Domain Specific Language (DSL). The DSL ap
 proach combined with the separation of concerns allows science code to be 
 developed productively as the code is written "just like the maths". The m
 odel is also being developed to be portable and performant as PSyclone can
 , in principle, target different programming model for different architect
 ures. Moreover, transformations of the internal representation can be deve
 loped for performance optimisations. All three aspects of model developmen
 t will be demonstrated.
END:VEVENT
END:VCALENDAR

