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:20181221T160906Z
LOCATION:D161
DTSTART;TZID=America/Chicago:20181112T090000
DTEND;TZID=America/Chicago:20181112T173000
UID:submissions.supercomputing.org_SC18_sess158@linklings.com
SUMMARY:9th Workshop on Latest Advances in Scalable Algorithms for Large-S
cale Systems
DESCRIPTION:Workshop\nAlgorithms, Heterogeneous Systems, Resiliency, Works
hop Reg Pass\n\nIntroduction - 9th Workshop on Latest Advances in Scalable
Algorithms for Large-Scale Systems\n\nAlexandrov, Geist, Dongarra, Engelm
ann\n\nNovel scalable scientific algorithms are needed in order to enable
key science applications to exploit the computational power of large-scale
systems. This is especially true for the current tier of leading petascal
e machines and the road to exascale computing as HPC systems continue to s
cale up in ...\n\n---------------------\nCommunication Reduced Multi-Times
tep Algorithm for Real-Time Wind Simulation on GPU-Based Supercomputers\n\
nOnodera, Idomura, Ali, Shimokawabe\n\nWe develop a communication reduced
multi-time-step (CRMT) algorithm for a Lattice Boltzmann method (LBM) base
d on a block-structured adaptive mesh refinement (AMR). This algorithm is
based on the temporal blocking method, and can improve computational effic
iency by replacing a communication bottlene...\n\n---------------------\nD
ynamic Load Balancing of Plasma and Flow Simulations\n\nDiamond, Smith, Yo
on, Shephard\n\nExtracting performance from simulations with complex infor
mation dependencies\non massively parallel computers requires the comput
ational work to be evenly\ndistributed across the processing resources w
hile maintaining low\ncommunication costs.\nPlasma simulations using a
particle-in-cell method...\n\n---------------------\nShift-Collapse Accel
eration of Generalized Polarizable Reactive Molecular Dynamics for Machine
Learning-Assisted Computational Synthesis of Layered Materials\n\nLiu, Ti
wari, Sheng, Krishnamoorthy, Hong...\n\nReactive molecular dynamics is a p
owerful simulation method for describing chemical reactions. Here, we intr
oduce a new generalized polarizable reactive force-field (ReaxPQ+) model t
o significantly improve the accuracy by accommodating the reorganization o
f surrounding media. The increased computati...\n\n---------------------\n
Iterative Randomized Algorithms for Low Rank Approximation of Terascale Ma
trices with Small Spectral Gaps\n\nIyer, Gittens, Carothers, Drineas\n\nRa
ndomized approaches for low rank matrix approximations have become popular
in recent years and often offer significant advantages over classical alg
orithms because of their scalability and numerical robustness on distribut
ed memory platforms. We present a distributed implementation of randomized
b...\n\n---------------------\nCommunication Avoiding Multigrid Precondit
ioned Conjugate Gradient Method for Extreme Scale Multiphase CFD Simulatio
ns\n\nIdomura, Ina, Yamashita, Onodera, Yamada...\n\nA communication avoid
ing (CA) multigrid preconditioned conjugate gradient method (CAMGCG) is a
pplied to the pressure Poisson equation in a multiphase CFD code JUPITER,
and its computational performance and convergence property are compared ag
ainst CA Krylov methods. A new geometric multigrid precon...\n\n----------
-----------\nMachine Learning-Aided Numerical Linear Algebra: Convolution
al Neural Networks for the Efficient Preconditioner Generation\n\nGötz\n\n
Markus Götz received his Bachelors and Masters degree in Software Engineer
ing from the University of Potsdam in 2010 and 2014 respectively. Afterwar
ds, he has been with the Research Center Jülich and the University of Icel
and, from which Markus obtained his PhD degree in Computational Engineerin
g fo...\n\n---------------------\nA General-Purpose Hierarchical Mesh Part
itioning Method with Node Balancing Strategies for Large-Scale Numerical S
imulations\n\nKong, Stogner, Gaston, Peterson, Permann...\n\nLarge-scale p
arallel numerical simulations are essential for a wide range of engineerin
g problems\nthat involve complex, coupled physical processes interacting
across a broad range of spatial\nand temporal scales. The data structur
es involved in such simulations (meshes, sparse matrices, etc.) are...\n\n
---------------------\nNon-Collective Scalable Global Network Based on Loc
al Communications\n\nBerghoff, Kondov\n\nTo efficiently perform collective
communications in current high-performance computing systems is a time-co
nsuming task.\nWith future exascale systems, this communication time will
be increased further.\nHowever, global information is frequently required
in various physical models.\nBy exploiting domai...\n\n-------------------
--\nIntroduction\n\nAlexandrov\n\n---------------------\nKeynote 1: A Few
Scheduling Problems for Resilience at Scale\n\nRobert\n\nThe talk will add
ress scheduling problems related to multi-level checkpointing, IO interfer
ence,\nsilent errors detection and correction, and workflows.\n\n---------
------------\nKeynote 2: HPC and AI as Drivers for Industrial Engagement\n
\nKennedy\n\nThe Hartree Centre is transforming UK industry through high p
erformance computing, big data and cognitive technologies. Backed by over
£170 million of government funding and significant strategic partnerships
with organisations such as IBM and Atos, the Hartree Centre is home to som
e of the most tec...\n\n---------------------\nKeynote 3: Hierarchical Alg
orithms on Hierarchical Architectures\n\nKeyes\n\nSome algorithms achieve
optimal arithmetic complexity with low arithmetic intensity (flops/Byte),
or possess high arithmetic intensity but lack optimal complexity, while so
me hierarchical algorithms, such as Fast Multipole and its H-matrix algebr
aic generalizations, realize a combination of optimal ...\n\n-------------
--------\nWorkshop Morning Break\n\n\n\n---------------------\nWorkshop Af
ternoon Break\n\n\n\n---------------------\nWorkshop Lunch (on your own)\n
\n\n\n---------------------\nEvent-Triggered Communication in Parallel Com
puting\n\nGhosh, Saha, Gupta, Tryggvason\n\nCommunication overhead in para
llel systems can be a significant bottleneck in scaling up parallel comput
ation. In this paper, we propose event-triggered communication methods to
reduce such communication overhead for numerical simulation of partial dif
ferential equations. As opposed to traditional c...\n\n-------------------
--\nLow Thread-Count Gustavson: A Multithreaded Algorithm for Sparse Matri
x-Matrix Multiplication Using Perfect Hashing\n\nElliott, Siefert\n\nSpars
e matrix-matrix multiplication is a critical kernel for several scientific
computing applications, especially the setup phase of algebraic multigrid
. The MPI+X programming model, which is growing in popularity, requires th
at such kernels be implemented in a way that exploits on-node parallelism.
..\n\n---------------------\nOn Advanced Monte Carlo Methods for Linear Al
gebra on Advanced Accelerator Architectures\n\nLebedev, Alexandrov\n\nIn t
his paper we present computational experiments performed using the Markov
Chain Monte Carlo Matrix Inversion (MCMCMI) on several architectures of NV
IDIA accelerators and two iterations of the Intel x86 architecture and inv
estigate their impact on performance and scalability of the method.\nThe m
e...\n
END:VEVENT
END:VCALENDAR