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:20200129T163600Z
LOCATION:607
DTSTART;TZID=America/Denver:20191118T090000
DTEND;TZID=America/Denver:20191118T173000
UID:submissions.supercomputing.org_SC19_sess124@linklings.com
SUMMARY:10th Workshop on Latest Advances in Scalable Algorithms for Large-
Scale Systems
DESCRIPTION:Workshop\n\nNovel scalable scientific algorithms are needed in
order to enable key science applications to exploit the computational pow
er of large-scale systems. This is especially true for the current tier of
leading petascale and machines and the road to exascale computing as HPC
systems continue to scale up in compute node and processor core count. The
se extreme-scale systems require novel scientific algorithms to hide netwo
rk and memory latency, have very high computation/communication overlap, h
ave minimal communication, and have no synchronization points. With the ad
vent of Big Data and AI in the past few years, the need of such scalable m
athematical methods and algorithms able to handle data and compute intensi
ve applications at scale becomes even more important.\n\nScientific algori
thms for multi-petaflop and exaflop systems also need to be fault tolerant
and fault resilient, since the probability of faults increases with scale
. Resilience at the system software and at the algorithmic level is needed
as a crosscutting effort. Finally, with the advent of heterogeneous compu
te nodes that employ standard processors as well as GPGPUs, scientific alg
orithms need to match these architectures to extract the most performance.
This includes different system-specific levels of parallelism as well as
co-scheduling of computation. Key science applications require novel mathe
matics and mathematical models and system software that address the scalab
ility and resilience challenges of current- and future-generation extreme-
scale HPC systems.\n\nhttps://www.csm.ornl.gov/srt/conferences/Scala/2019\
n\nToward Half-Precision Computation for Complex Matrices: A Case Study fo
r Mixed Precision Solvers on GPUs\n\nAbdelfattah, Tomov, Dongarra\n\nLow-p
recision computations are popular in machine learning and artificial intel
ligence (AI) applications. Hardware architectures, such as high-end GPUs,
now support native 16-bit floating point arithmetic (i.e. half-precision).
While half-precision provides a natural 2x/4x speedups against the perf.
..\n\n---------------------\nGeneric Matrix Multiplication for Multi-GPU A
ccelerated Distributed-Memory Platforms Over PaRSEC\n\nHerault, Robert, Bo
silca, Dongarra\n\nThis paper introduces a generic and flexible matrix-mat
rix\nmultiplication algorithm $C = A \times B$ for state-of-the-art\nc
omputing platforms. Typically, these platforms are\ndistributed-memory m
achines whose nodes are equipped with several\naccelerators (e.g., 6 GPU
s per node for Summit. To\n...\n\n---------------------\nGPU Acceleration
of Communication Avoiding Chebyshev Basis Conjugate Gradient Solver for M
ultiphase CFD Simulations\n\nAli, Onodera, Idomura, Ina, Imamura\n\nIterat
ive methods for solving large linear systems are common parts of computati
onal fluid dynamics (CFD) codes. The Preconditioned Conjugate Gradient (P-
CG) method is one of the most widely used iterative methods. However, in t
he P-CG method, global collective communication is a crucial bottleneck ..
.\n\n---------------------\nParallel SFC-Based Mesh Partitioning and Load
Balancing\n\nBorrell, Oyarzun, Dosimont, Houzeaux\n\nModern supercomputers
allow the simulation of complex phenomena with increased accuracy. Eventu
ally, this requires finer geometric discretizations with larger numbers of
mesh elements. In this context, and extrapolating to the Exascale paradig
m, meshing operations such as generation, adaptation or p...\n\n----------
-----------\n10th Workshop on Latest Advances in Scalable Algorithms for L
arge-Scale Systems\n\nAlexandrov, Dongarra, Geist, Engelmann\n\nNovel scal
able scientific algorithms are needed in order to enable key science appli
cations to exploit the computational power of large-scale systems. This is
especially true for the current tier of leading petascale and machines an
d the road to exascale computing as HPC systems continue to scale up...\n\
n---------------------\nOptimization of a Solver for Computational Materia
ls and Structures Problems on NVIDIA Volta and AMD Instinct GPUs\n\nZubair
, Warner, Wagner\n\nThe Scalable Implementation of Finite Elements by NASA
(ScIFEN) is a software package developed to solve complex computational m
aterials and structures problems using the finite element method (FEM). In
this paper, we describe optimization techniques to speed up the linear so
lver computation that oc...\n\n---------------------\nMaking Speculative S
cheduling Robust to Incomplete Data\n\nGainaru, Pallez\n\nWe study in this
work the robustness of Speculative Scheduling to the\nincompleteness of d
ata. Speculative scheduling has been introduce as a solution\nto incorpora
te future types of applications into the design of HPC schedulers, specifi
cally\napplications whose runtime is not perfectly known but can ...\n\n--
-------------------\nToward Accelerated Unstructured Mesh Particle-in-Cell
\n\nDiamond, Smith, Shephard\n\nThe effective execution of unstructured me
sh based particle-in-cell, PIC, simulations on\nGPUs requires careful de
sign and implementation choices to ensure performance\nwhile maintaining
productive programmability.\nThis paper overviews the developing PUMIPi
c library\nthat employs a set of mesh ...\n\n---------------------\nAfte
rnoon Break\n\n\n\n---------------------\nLunch Break\n\n\n\n-------------
--------\nMorning Break\n\n\n\n---------------------\nKeynote 3: The Extre
me-Scale Scientific Software Stack and Its Promise for the Exascale Comput
ing Era\n\nHeroux\n\nOpen source, community-developed reusable scientific
software represents a large and growing body of capabilities. Linux distr
ibutions, vendor software stacks and individual disciplined software produ
ct teams provide the scientific computing community with usable holistic s
oftware environments cont...\n\n---------------------\nParallel Multigrid
Methods on Manycore Clusters with IHK/McKernel\n\nNakajima, Gerofi, Ishika
wa, Horikoshi\n\nThe parallel multigrid method is expected to play an impo
rtant role in large-scale scientific computing on exa-scale supercomputer
systems. Previously we proposed Hierarchical Coarse Grid Aggregation (hCGA
), which dramatically improved the performance of the parallel multigrid s
olver when the number...\n\n---------------------\nKeynote 1: Exascale App
lication Progress and Challenges\n\nKothe, Diachin\n\nFor exascale applica
tions under development in the U.S Department of Energy (DOE) Exascale Com
puting Project (ECP), they are charged with delivering comprehensive scien
ce-based computational applications that effectively exploit exascale HPC
technologies to provide breakthrough modeling and simulati...\n\n---------
------------\nKeynote 2: Toward Scaling Deep Learning to 100,000 Processor
s - The Fugaku Challenge\n\nMatsuoka\n\nModern AI with deep learning poses
significant overhead in training over very large data sets, whereby the u
se of HPC techniques to compute in parallel on a large machine is becoming
increasingly popular. However, most of the efforts have been on GPUs at r
elatively low scale, in the order of a few h...\n\n---------------------\n
Extreme Scale Phase-Field Simulation of Sintering Processes\n\nHierl, Hötz
er, Seiz, Reiter, Nestler\n\nThe sintering process, which turns loose powd
ers into dense materials, is naturally found in the formation of glaciers,
but is also the indispensable process to manufacture ceramic materials.\n
The dynamically evolving microstructure, which is established during this
process, largely influences the re...\n\n\nTag: Workshop Reg Pass, Algorit
hms, Scalable Computing\n\nRegistration Category: Workshop Reg Pass, Algor
ithms, Scalable Computing
END:VEVENT
END:VCALENDAR