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:20260422T143140Z
LOCATION:B304
DTSTART;TZID=America/New_York:20241117T144500
DTEND;TZID=America/New_York:20241117T150000
UID:submissions.supercomputing.org_SC24_sess745_ws_scalah102@linklings.com
SUMMARY:High-Performance EigenSolver Combining EigenExa and Iterative Refi
 nement
DESCRIPTION:Yuki Uchino and Toshiyuki Imamura (RIKEN)\n\nThis study propos
 es a high-performance and reliable eigensolver via mixed-precision arithme
 tic between ordinary and highly-accurate precisions. Eigenvalue decomposit
 ion is ubiquitous in simulations. Various eigensolvers for computing appro
 ximations have been developed thus far. If eigenvalues are narrowly cluste
 red, the computation of eigenvectors may be ill-posed. Thus, the computed 
 eigenpairs may not be sufficiently accurate and lack reliability. In this 
 study, we introduce mixed-precision iterative refinement methods to improv
 e the accuracy of eigenvectors obtained using numerical methods. This appr
 oach contributes to obtaining sufficiently accurate results without arbitr
 ary precision eigensolvers. We construct a high-performance and reliable e
 igensolver by combining the iterative refinement methods and EigenExa, a m
 odern high-performance solver for large-scale and highly parallel computat
 ions. Numerical experiment results demonstrate the accuracy of the results
  and performance benchmark of the proposed approach.\n\nTag: Algorithms, H
 eterogeneous Computing\n\nRegistration Category: Workshop Reg Pass\n\nSess
 ion Chairs: Vassil Alexandrov (Hartree Centre, STFC); Jack Dongarra (Unive
 rsity of Tennessee, Knoxville; Oak Ridge National Laboratory (ORNL)); Erik
  Draeger (Lawrence Livermore National Laboratory (LLNL), Center for Applie
 d Scientific Computing); Christian Engelmann (Oak Ridge National Laborator
 y (ORNL)); and Dieter A. Kranzlmueller (Ludwig-Maxmilians-Universität Münc
 hen, Leibniz Supercomputing Centre (LRZ))\n\n
END:VEVENT
END:VCALENDAR