BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260202T201805Z
LOCATION:274
DTSTART;TZID=America/Chicago:20251116T153000
DTEND;TZID=America/Chicago:20251116T155000
UID:submissions.supercomputing.org_SC25_sess201_ws_memo101@linklings.com
SUMMARY:Umpire: Portable Memory Management for High-Performance Computing 
 Applications
DESCRIPTION:Kristi Belcher (Lawrence Livermore National Laboratory) and Da
 vid Beckingsale (Lawrence Livermore National Laboratory (LLNL))\n\nModern 
 high-performance computing (HPC) systems present application\ndevelopers w
 ith increasingly complex memory hierarchies\nthat include multiple types o
 f memory with varying access patterns,\ncapacities, and performance charac
 teristics. Managing these\nresources efficiently while maintaining code po
 rtability across different\narchitectures remains a significant challenge.
  To address\nthese challenges, Umpire was developed at Lawrence Livermore\
 nNational Laboratory (LLNL) as an open-source library that provides\na uni
 fied, portable memory management API for modern HPC platforms\nwith multip
 le memory devices like NUMA and GPUs. This\npaper explores Umpire’s design
  principles, outlines Umpire’s primary\nperformance advantages, and examin
 es howits memory pools\ncan provide speedups of 15x or greater. Next, it d
 emonstrates how\nits integration with the RAJA Portability Suite enables t
 he development\nof portable and performant HPC applications. With real-wor
 ld\nexamples from LLNL’s production codes, Umpire provides a comprehensive
 \nsolution for managing the challenges of performance\nportability in mode
 rn HPC environments.\n\nRecording: Livestreamed, Recorded\n\nRegistration 
 Category: Technical Program Reg Pass, Workshop Reg Pass\n\nSession Chairs:
  Stephen L. Olivier (Sandia National Laboratories), Maya Gokhale (Lawrence
  Livermore National Laboratory (LLNL)), Ivy Peng (KTH Royal Institute of T
 echnology), Kyle Hale (Oregon State University), and Ronald Minnich (Hewle
 tt Packard Enterprise (HPE))\n\n
END:VEVENT
END:VCALENDAR
