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:20260422T000604Z
LOCATION:DEF Concourse
DTSTART;TZID=America/Denver:20231115T100000
DTEND;TZID=America/Denver:20231115T170000
UID:submissions.supercomputing.org_SC23_sess303_rpost170@linklings.com
SUMMARY:Simulating Application Agnostic Process Assignment for Graph Workl
 oads on Dragonfly and Fat Tree Topologies
DESCRIPTION:Md Nahid Newaz (Oakland University); Sayan Ghosh, Joshua Suett
 erlein, and Nathan Tallent (Pacific Northwest National Laboratory (PNNL));
  and Hua Ming (Oakland University)\n\nDistributed-memory graph application
 s are dominated by communication and synchronization overheads. For such a
 pplications, the communication pattern comprises of variable-sized data ex
 changes between process neighbors in a process graph topology, which unlik
 e process grid for rectangular problems is difficult to optimize for enhan
 cing the locality in a sustainable fashion.\n\nProcess assignment or remap
 ping can improve the communication performance, however, existing solution
 s mostly caters to Cartesian process topologies and not the graph topology
 . In this work, we propose application and topology agnostic process remap
 ping strategies for graph applications. For two communication intensive di
 stributed-memory graph applications (graph clustering and triangle countin
 g), we demonstrate about 30% improvements in the overall execution times t
 hrough various remapping methodologies via SST-based packet-level simulati
 ons on Dragonfly and Fat Tree based network topologies.\n\nTag: Artificial
  Intelligence/Machine Learning, Architecture and Networks, Heterogeneous C
 omputing, I/O and File Systems, Performance Measurement, Modeling, and Too
 ls, Post-Moore Computing, Programming Frameworks and System Software, Quan
 tum Computing\n\nRegistration Category: Tech Program Reg Pass, Exhibits Re
 g Pass\n\n
END:VEVENT
END:VCALENDAR