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:20211207T055402Z
LOCATION:Second Floor Atrium
DTSTART;TZID=America/Chicago:20211116T083000
DTEND;TZID=America/Chicago:20211116T170000
UID:submissions.supercomputing.org_SC21_sess256_drs104@linklings.com
SUMMARY:Operator Splittings on GPUs
DESCRIPTION:Doctoral Showcase, Posters\n\nOperator Splittings on GPUs\n\nK
 lein, Strzodka\n\nOperator splittings are a successful method for the solu
 tion of parabolic partial differential equations. In this thesis, the same
  ideas are applied for general sparse linear equation systems. For general
  graphs, which are induced by the sparse matrix of the equation system, a 
 parallel segmentation algorithm is used to extract one-dimensional segment
 s, and create appropriate renumberings, such that the permuted subgraph ha
 s a tridiagonal form. With these tridiagonal factors, multiplicative and a
 lternating operator splittings are created and applied as preconditioners 
 in iterative solvers. The presented experiments with matrices from the Spa
 rse Matrix Collection show that the proposed preconditioners can achieve b
 etter convergence rate and performance than state-of-the art preconditione
 rs, like the ILU-ISAI implemented in the Magma library. The tridiagonal fa
 ctors are solved with the Recursive Partitioned Tridiagonal Schur Compleme
 nt Algorithm (RPTS), which was developed and implemented within the scope 
 of this thesis. RPTS is a hierarchically tridiagonal GPU solver with scale
 d partial pivoting, which runs at maximum GPU memory bandwidth, and outper
 forms the numerically stable tridiagonal solver of cuSPARSE by approximate
 ly factor 5 for large problem sizes.\n\nTag: In-Person Only\n\nRegistratio
 n Category: Tech Program Reg Pass, Exhibit Hall Only
END:VEVENT
END:VCALENDAR
