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:20260422T000712Z
LOCATION:301-302-303
DTSTART;TZID=America/Denver:20231115T163000
DTEND;TZID=America/Denver:20231115T170000
UID:submissions.supercomputing.org_SC23_sess181_pap609@linklings.com
SUMMARY:A Quantitative Approach for Adopting Disaggregated Memory in HPC S
 ystems
DESCRIPTION:Jacob Wahlgren and Gabin Schieffer (KTH Royal Institute of Tec
 hnology, Sweden); Maya Gokhale (Lawrence Livermore National Laboratory (LL
 NL)); and Ivy B. Peng (KTH Royal Institute of Technology, Sweden)\n\nMemor
 y disaggregation has recently been adopted in major data centers to improv
 e resource utilization, driven by cost and sustainability. Meanwhile, stud
 ies on large-scale HPC facilities have also highlighted memory under-utili
 zation. A promising and non-disruptive option for memory disaggregation is
  rack-scale memory pooling, where node-local memory is supplemented by sha
 red memory pools. This work outlines the prospects and requirements for ad
 option and clarifies several misconceptions. We propose a quantitative met
 hod for dissecting application requirements on the memory system in three 
 levels, moving from general, to multi-tier memory, and then to memory pool
 ing. We also provide tools to facilitate the quantitative approach. We eva
 luated a set of representative HPC workloads on an emulated platform. Our 
 results show that interference in memory pooling has varied application im
 pact, depending on access ratio and arithmetic intensity. Finally, our met
 hod is applied in two case studies to show benefits at both the applicatio
 n and system level.\n\nTag: Cloud Computing, Distributed Computing, Data M
 ovement and Memory, Performance Measurement, Modeling, and Tools\n\nRegist
 ration Category: Tech Program Reg Pass\n\nReproducibility Badges: Artifact
  Available, Artifact Functional, Results Reproduced\n\nSession Chair: Jay 
 Lofstead (Sandia National Laboratories, University of New Mexico)\n\n
END:VEVENT
END:VCALENDAR