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:20260522T150109Z
LOCATION:C2/3/4 Ballroom
DTSTART;TZID=America/Chicago:20181113T083000
DTEND;TZID=America/Chicago:20181113T170000
UID:submissions.supercomputing.org_SC18_sess340_drs115@linklings.com
SUMMARY:Scalable Non-Blocking Krylov Solvers for Extreme-Scale Computing
DESCRIPTION:Paul R. Eller and William Gropp (University of Illinois)\n\nTh
 is study investigates preconditioned conjugate gradient method variations 
 designed to reduce communication costs by decreasing the number of allredu
 ces and overlapping communication with computation using a non-blocking al
 lreduce. Experiments show scalable PCG methods can outperform standard PCG
  at scale and demonstrate the robustness of these methods.<br /><br />To d
 evelop the most optimal Krylov methods we need a clear understanding of th
 e factors limiting performance at scale. Detailed timings and network coun
 ters are used to more thoroughly measure the performance of these methods.
  Performance models with penalty terms are developed that provide reasonab
 le explanations of observed performance and guide development of optimizat
 ions. The effectiveness of scalable PCG methods and these performance anal
 ysis tools is demonstrated using Quda and Nek5000, two HPC applications se
 eking improved performance at scale.\n\nRegistration Category: Workshop Re
 g Pass, Tutorial Reg Pass, Tech Program Reg Pass, Exhibits Reg Pass, Exhib
 its - Exhibit Hall Only Reg Pass\n\n
END:VEVENT
END:VCALENDAR