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 F 1
DTSTART;TZID=Europe/Stockholm:20190613T174500
DTEND;TZID=Europe/Stockholm:20190613T181500
UID:submissions.pasc-conference.org_PASC19_sess140_msa230@linklings.com
SUMMARY:FLASH Software Design for Code Longevity
DESCRIPTION:Minisymposium\nComputer Science and Applied Mathematics, Emerg
 ing Application Domains, Physics, Engineering\n\nFLASH Software Design for
  Code Longevity\n\nDubey, Weide, O'Neal, Chawdhary\n\nThe Exascale Computi
 ng Project coincides with rapid hardware architecture changes. The normal 
 high-performance computing (HPC) model of targeted optimization for specif
 ic platforms has been somewhat impractical for large multiphysics codes al
 l along, it has now become impossible. Such codes can take multiple years 
 to refactor, and with rapidly changing platforms they could get caught in 
 an endless cycle of continuously refactoring. The only safeguard against t
 his scenario is to design codes with an architecture that allows for separ
 ation of concerns along with a hierarchy of abstractions. The FLASH code h
 as always taken these two principles into its design consideration, though
  in the past, concerns were few and hierarchy was short. For the upcoming 
 exascale platforms, there are many more layers in the hierarchy and many m
 ore concerns that need to be accounted for. Under the ECP project, we are 
 redesigning the FLASH code architecture to meet these requirements. Additi
 onally, because exascale will not be the end of computing with FLASH, our 
 design is by necessity more forward-looking and general. Here, we describe
  our component based design approach that we hope will stand the test of t
 ime.
END:VEVENT
END:VCALENDAR