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:20260522T150116Z
LOCATION:D165
DTSTART;TZID=America/Chicago:20181112T090000
DTEND;TZID=America/Chicago:20181112T093000
UID:submissions.supercomputing.org_SC18_sess161_ws_pmbsf103@linklings.com
SUMMARY:Improving MPI Reduction Performance for Manycore Architectures wit
 h OpenMP and Data Compression
DESCRIPTION:Hongzhang Shan and Samuel Williams (Lawrence Berkeley National
  Laboratory) and Calvin Johnson (San Diego State University)\n\nMPI reduct
 ions are widely used in many scientific applications and often become the 
 scaling performance bottleneck. When performing reductions on vectors, dif
 ferent algorithms have been developed to balance messaging overhead and ba
 ndwidth. However, most implementations have ignored the effect of single-t
 hread performance not scaling as fast as aggregate network bandwidth. In t
 his work, we propose, implement, and evaluate two approaches (threading an
 d exploitation of sparsity) to accelerate MPI reductions on large vectors 
 when running on manycore-based supercomputers. Our benchmark results show 
 that our new techniques improve the MPI_Reduce performance up to 4x and im
 prove BIGSTICK application performance by up to 2.6x.\n\nTag: Benchmarks, 
 Parallel Programming Languages, Libraries, and Models, Performance, Simula
 tion\n\nRegistration Category: Workshop Reg Pass\n\nSession Chair: Steven 
 A. Wright (University of York, England)\n\n
END:VEVENT
END:VCALENDAR