BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T143140Z
LOCATION:B308
DTSTART;TZID=America/New_York:20241121T160000
DTEND;TZID=America/New_York:20241121T163000
UID:submissions.supercomputing.org_SC24_sess371_pap244@linklings.com
SUMMARY:Doubling Graph Traversal Efficiency to 198 TeraTEPS on the Superco
 mputer Fugaku
DESCRIPTION:Junya Arai (Nippon Telegraph and Telephone Corporation); Masah
 iro Nakao (RIKEN); Yuto Inoue, Kanto Teranishi, and Koji Ueno (Fixstars Co
 rporation); Keiichiro Yamamura (Institute of Science Tokyo); Mitsuhisa Sat
 o (RIKEN); and Katsuki Fujisawa (Institute of Science Tokyo)\n\nBreadth-fi
 rst search (BFS) is a fundamental building block of various high-performan
 ce computing applications beyond graph analysis and also known as a benchm
 ark problem in the Graph500 list. The increasing volume of global data dem
 ands efficient distributed BFS, which, however, is hindered by the high co
 mmunication costs of exchanging vertex data between compute nodes. To addr
 ess this challenge, this paper introduces four techniques: (i) forest prun
 ing, which reduces the number of vertices by eliminating those unnecessary
  for the search; (ii) group reordering and (iii) multilevel bitmap compres
 sion, which decrease the memory footprint of graph data, thereby enabling 
 fewer nodes to manage larger graphs; and (iv) adaptive parameter tuning, w
 hich quickly optimizes the hyperparameters of the BFS algorithm. In the ev
 aluation using 152,064 nodes of the supercomputer Fugaku, our implementati
 on achieved 198 tera-traversed edges per second, doubling the performance 
 reported in the latest Graph500-related study on Fugaku.\n\nTag: Accelerat
 ors, Algorithms, Data Movement and Memory, Graph Algorithms\n\nRegistratio
 n Category: Tech Program Reg Pass\n\nSession Chair: Helen Xu (Georgia Inst
 itute of Technology)\n\n
END:VEVENT
END:VCALENDAR