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:20260522T150115Z
LOCATION:D166
DTSTART;TZID=America/Chicago:20181111T113000
DTEND;TZID=America/Chicago:20181111T120000
UID:submissions.supercomputing.org_SC18_sess174_ws_exampi101@linklings.com
SUMMARY:Optimal Algorithms for Half-Duplex Inter-Group All-to-All Broadcas
 t on Fully Connected and Ring Topologies
DESCRIPTION:Qiao Kang, Alok Choudhary, Ankit Agrawal, and Wei-keng Liao (N
 orthwestern University)\n\nHalf-duplex inter-group collective communicatio
 ns are bipartite message transfer patterns such that the processes in a se
 nder group pass messages to the processes in a receiver group. These commu
 nication patterns serve as basic operations for scientific application wor
 kflows. In this paper, we present optimal parallel algorithms for half-dup
 lex inter-group all-to-all broadcast under bidirectional communication con
 straint on fully connected and ring topologies. We implement the algorithm
 s using MPI communication functions and perform experiments on Cori. For t
 he fully connected topology case, we compare our algorithms with productio
 n MPI libraries. For the ring topology case, we implement our proposed alg
 orithms using MPI\_Sendrecv function to emulate a ring topology environmen
 t. The proposed algorithms are compared with the intra-group Allgather alg
 orithm emulated under the same environment. Message sizes ranging from 32K
 B to 4MB are used for evaluations. The proposed algorithms for fully conne
 cted topology are up to 5 times faster than the root gathering algorithm a
 dopted by MPICH. The proposed algorithms for the ring topology are up to 1
 .4 times faster than the intra-group Allgather algorithm.\n\nTag: Exascale
 , MPI, Networks, System Software\n\nRegistration Category: Workshop Reg Pa
 ss\n\nSession Chairs: Purushotham Bangalore (University of Alabama); Ryan 
 Grant (Queen's University, Canada; Power API); Daniel Holmes (Collis-Holme
 s Innovations); and Anthony Skjellum (Tennessee Technological University, 
 ASCEND-TNTECH)\n\n
END:VEVENT
END:VCALENDAR
