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:20260924T172458Z
LOCATION:B308
DTSTART;TZID=America/New_York:20241121T153000
DTEND;TZID=America/New_York:20241121T160000
UID:submissions.supercomputing.org_SC24_sess371_pap242@linklings.com
SUMMARY:Enumeration of Billions of Maximal Bicliques in Bipartite Graphs w
 ithout Using GPUs
DESCRIPTION:Zhe Pan (Zhejiang University, China; Hangzhou High-Tech Zone (
 Binjiang) Institute of Blockchain and Data Security); Shuibing He and Xu L
 i (Zhejiang University, China); Xuechen Zhang (Washington State University
  Vancouver); Yanlong Yin (Zhejiang University, China); Rui Wang and Lidan 
 Shou (Zhejiang University, China; Hangzhou High-Tech Zone (Binjiang) Insti
 tute of Blockchain and Data Security); Mingli Song (Zhejiang University, C
 hina); Xian-He Sun (Illinois Institute of Technology); and Gang Chen (Zhej
 iang University, China)\n\nMaximal biclique enumeration (MBE) is crucial i
 n bipartite graph analysis. Recent studies rely on extensive set intersect
 ions on static bipartite graphs to solve the MBE problem. However, the com
 putational subgraphs dynamically change during enumeration, leading to red
 undant memory accesses and degraded set intersection performance. To overc
 ome this limitation, we propose an AdaMBE algorithm. First, we redesign it
 s core operations using local neighborhood information derived from comput
 ational subgraphs to minimize redundant memory accesses. Second, we dynami
 cally create computational subgraphs using bitmaps leveraging its fast bit
 wise operations to accelerate set intersections. Finally, we integrate the
 m in AdaMBE. Our experimental results show that AdaMBE is 1.6×−49.7× faste
 r than its closest CPU-based competitor and successfully enumerates all 19
  billion maximal bicliques on the TVTropes dataset, a large task beyond th
 e capabilities of existing algorithms. Notably, on certain datasets, our p
 arallel version, ParAdaMBE, on CPUs even outperforms GMBE on GPUs by up to
  5.07×.\n\nTag: Accelerators, Algorithms, Data Movement and Memory, Graph 
 Algorithms\n\nRegistration Category: Tech Program Reg Pass\n\nSession Chai
 r: Helen Xu (Georgia Institute of Technology)\n\n
END:VEVENT
END:VCALENDAR
