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:20210916T132448Z
LOCATION:Lise Girardin
DTSTART;TZID=Europe/Stockholm:20210705T170000
DTEND;TZID=Europe/Stockholm:20210705T173000
UID:submissions.pasc-conference.org_PASC21_sess117_msa206@linklings.com
SUMMARY:​Programming the Superfluid DFT on Leadership-Class Supercom
 puters
DESCRIPTION:Minisymposium\n\n​Programming the Superfluid DFT on Lead
 ership-Class Supercomputers\n\nRoche, Bulgac, Magierski, Sekizawa, Wlazlow
 ski\n\nDeveloping or migrating application codes to effectively utilize pe
 ta- and exa- scale computer architectures poses interesting and difficult 
 challenges to programmers. A specific example is preparing codes for accel
 erator-based hybrid architectures. Here, despite software tools that aim t
 o simplify or automate the process, experience suggests that to achieve hi
 gh performance or scalability often requires distinct or restructured algo
 rithms, data structures, and programming or execution techniques be used. 
 I will describe the parallel discretization strategy of the stationary and
  time-dependent superfluid DFT theory, will review the scaling and perform
 ance status of the production codes for some of today's most powerful supe
 rcomputers, and discuss adjustments required to successfully target some e
 xpected architectures. This work has been used as a case study within the 
 Application Assessment project of the United States Department of Energy E
 xascale Computing Project.\n\nDomain: CS and Math, Physics
END:VEVENT
END:VCALENDAR