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:20211207T054805Z
LOCATION:230-231-232
DTSTART;TZID=America/Chicago:20211117T113000
DTEND;TZID=America/Chicago:20211117T120000
UID:submissions.supercomputing.org_SC21_sess168_pap214@linklings.com
SUMMARY:Krill: A Compiler and Runtime System for Concurrent Graph Processi
 ng
DESCRIPTION:Paper\n\nKrill: A Compiler and Runtime System for Concurrent G
 raph Processing\n\nChen, Shen, Xiao, Lu\n\nAs a large number of emerging g
 raph applications spread across different domains, the need for processing
  massive concurrent graph jobs (CGJs) is increasing. However, existing gra
 ph processing systems designed for a single job cannot efficiently tackle 
 multiple CGJs, where they suffer from interfering memory access patterns a
 nd inefficient property management. In this paper, we introduce Krill, a c
 ompiler and runtime system for processing concurrent graph jobs. In the co
 mpiler, we propose leveraging the property buffer to easily write and mana
 ge property data. In the runtime system, we propose a novel technique name
 d graph kernel fusion to reduce memory accesses, which fuses all the jobs 
 and processes them as a whole. Experimental results show our system signif
 icantly reduces the number of memory accesses for CGJs by more than 6x com
 pared with the baseline, and achieves up to 7.67x speedup with 3.84x short
 er response latency compared with GraphM, the state-of-the-art graph syste
 m.\n\nTag: Reproducibility Badge, Machine Learning and Artificial Intellig
 ence, Numerical Algorithms\n\nRegistration Category: Tech Program Reg Pass
 \n\nReproducibility Badges: Artifact Available, Artifact Functional, Resul
 ts Reproduced
END:VEVENT
END:VCALENDAR