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:20200129T163603Z
LOCATION:406
DTSTART;TZID=America/Denver:20191118T083000
DTEND;TZID=America/Denver:20191118T120000
UID:submissions.supercomputing.org_SC19_sess182_tut153@linklings.com
SUMMARY:Fast Parallel Direct Linear Solvers and Preconditioners
DESCRIPTION:Tutorial\n\nFast Parallel Direct Linear Solvers and Preconditi
oners\n\nGhysels, Li, Liu\n\nMatrix factorizations and the accompanying so
lution algorithms (e.g., triangular solution associated with the LU factor
ization) are often the most robust algorithmic choices for linear systems
from multi-physics and multi-scale simulations. They are indispensable too
ls for building various algebraic equation solvers. They can be used as di
rect solvers, as coarse-grid solvers in multigrid, or as preconditioners f
or iterative solvers. As we are approaching the exascale computing era, de
mand for algorithm innovation is increasingly high. It is imperative to de
velop optimal-complexity scalable algorithms both in flop count and more i
mportantly in data movement, such as, in the form of communication-avoidin
g formulations, and low-rank and butterfly compressions. On the software a
nd implementation side, it is imperative to exploit multiple levels of par
allelism presented by the heterogeneous node architectures through well or
chestrated use of MPI, OpenMP and GPU programming like CUDA. \n\nIn this t
utorial, we will present our recently developed novel techniques to addres
s scalability gaps. We will demonstrate their efficacies through three sol
ver libraries: SuperLU, STRUMPACK and ButterflyPACK, with representative u
se cases from simulations and data analytics. Through hands-on exercises,
the participants will learn how to use each solver most effectively for th
eir target problems and parallel machines.\n\nTag: Tutorial Reg Pass, Algo
rithms, Linear Algebra, Scalable Computing\n\nRegistration Category: Tutor
ial Reg Pass, Algorithms, Linear Algebra, Scalable Computing
URL:https://sc19.supercomputing.org/presentation/?id=tut153&sess=sess182
END:VEVENT
END:VCALENDAR