BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T000713Z
LOCATION:710
DTSTART;TZID=America/Denver:20231113T121000
DTEND;TZID=America/Denver:20231113T123000
UID:submissions.supercomputing.org_SC23_sess445_ws_scalah103@linklings.com
SUMMARY:Advancing the Distributed Multi-GPU ChASE Library through Algorith
 m Optimization and NCCL Library
DESCRIPTION:Xinzhe Wu and Edoardo Di Napoli (Jülich Supercomputing Centre)
 \n\nAs supercomputers become larger with powerful Graphics Processing Unit
  (GPU), traditional direct eigensolvers struggle to keep up with the hardw
 are evolution and scale efficiently due to communication and synchronizati
 on demands. Subspace eigensolvers, like the Chebyshev Accelerated Subspace
  Eigensolver (ChASE), have a simpler structure and can overcome communicat
 ion and synchronization bottlenecks. ChASE is a modern subspace eigensolve
 r that uses Chebyshev polynomials to accelerate the computation of extrema
 l eigenpairs of dense Hermitian eigenproblems. In this work we show how we
  have modified ChASE by rethinking its memory layout, introducing a novel 
 parallelization scheme, switching to a more performing communication-avoid
 ing algorithm for one of its inner module, and substituting MPI library by
  vendor-optimized NCCL library. The resulting library can tackle dense pro
 blems with size up to N=O(10^6), and scales effortlessly up to the full 90
 0 nodes---each one powered by 4xA100 NVIDIA GPUs---of the JUWELS Booster h
 osted at the Jülich Supercomputing Centre.\n\nTag: Algorithms, Heterogeneo
 us Computing, Large Scale Systems\n\nRegistration Category: Workshop Reg P
 ass\n\nSession Chairs: Vassil Alexandrov (Hartree Centre, STFC); Jack Dong
 arra (University of Tennessee, Knoxville; Oak Ridge National Laboratory (O
 RNL)); Christian Engelmann (Oak Ridge National Laboratory (ORNL)); Al Geis
 t (Oak Ridge National Laboratory (ORNL)); and Dieter A. Kranzlmueller (Lud
 wig-Maxmilians-Universität München, Leibniz Supercomputing Centre (LRZ))\n
 \n
END:VEVENT
END:VCALENDAR