BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20240116T185911Z
LOCATION:E Concourse
DTSTART;TZID=America/Denver:20231114T100000
DTEND;TZID=America/Denver:20231114T170000
UID:submissions.supercomputing.org_SC23_sess290_drs110@linklings.com
SUMMARY:Modernizing Simulation Software for the Exascale Era
DESCRIPTION:Doctoral Showcase, Posters\n\nNigel P. Tan (University of Tenn
 essee)\n\nModern HPC hardware is becoming increasingly heterogeneous and d
 iverse in the exascale era. The diversity of hardware and software stacks 
 adds additional development challenges to high performance simulations. On
 e common development approach is to re-engineer the code for each new targ
 et architecture in order to maximize performance. However, this re-enginee
 ring effort is no longer practical due to increasing heterogeneous hardwar
 e. Adding support for a single family of GPUs alone poses a significant ch
 allenge. Supporting each major vendor's hardware and software stacks takes
  valuable developer time away from optimizing and enhancing simulation cap
 abilities. Moving forward, the community must modernize the code developme
 nt process in order to achieve the greatest scientific output.\n\nIn this 
 work, we examine the challenges posed by emerging heterogeneous hardware. 
 These challenges include developing performance portable code, leveraging 
 hardware features targeting AI/ML for HPC applications, and difficulties m
 anaging limited I/O resources while checkpointing. To address these challe
 nges we present a modernization approach for scientific software that ensu
 res the following. Attain high performance and portability across architec
 tures using the Kokkos portability framework in addition to optimizations 
 to memory layout, sorting algorithms, and vectorization. Leverage alternat
 ive number formats such as half-precision and fixed-point to maximize usag
 e of the limited memory on GPUs and enable larger simulations. Reduce IO o
 verhead and storage requirements through the identification and eliminatio
 n of spatial-temporal redundancy in application data.\n\nTag: Heterogeneou
 s Computing, Software Engineering\n\nRegistration Category: Tech Program R
 eg Pass, Exhibits Reg Pass
END:VEVENT
END:VCALENDAR