BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Coupling a temperature field to level set advection in a Stefan ph
ase change problem
DTSTART;VALUE=DATE-TIME:20180723T092000Z
DTEND;VALUE=DATE-TIME:20180723T094000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-11@cern.ch
DESCRIPTION:Speakers: Mr. MCGOVERN\, Sean (Forschungszentrum Juelich)\nGiv
en the context of distributed parallelism\, the purpose of this contributi
on is to consider a solution procedure to the classical Stefan phase chang
e problem\, an example of a free boundary problem. In the Stefan problem\,
the interface between water and ice moves through a domain as the ice mel
ts due to a heat source. As an example of an interface problem\, we choose
to apply the level set method to implicitly track the evolution of the ic
e/water interface. The diffusive temperature equation must be solved over
the whole domain while an evolving interface separates regions endowed wi
th different material properties\, i.e. specific heat of water and ice. Th
e Gibbs-Thomson relation gives the discontinuity in the temperature field
at the interface and the interface velocity is given by the jump in the he
at flux. \n\nTo solve this kind of coupled nonlinear problem\, we can 1) e
volve the interface and 2) solve the temperature equation iteratively. Th
e first is a hyperbolic advection equation and the second is parabolic. As
is well-known\, these require different solution tactics. We consider a s
etup for which an analytical solution exists.\nThe approach remains pedago
gical as I would like to package this as a possible tutorial step.\n\nhttp
s://indico.sissa.it/event/23/session/2/contribution/11
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/11
END:VEVENT
BEGIN:VEVENT
SUMMARY:CFL: A domain-specific language for simplifying integration kernel
s
DTSTART;VALUE=DATE-TIME:20180724T074500Z
DTEND;VALUE=DATE-TIME:20180724T083000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-10@cern.ch
DESCRIPTION:Speakers: ARNDT\, Daniel (University Heidelberg)\nThere are se
veral different ways of describing bilinear forms in `deal.II`. All of the
se require to describe the assembly loop inside the program. This talk dis
cusses an approach for providing the assembly operator from the outside us
ing a domain-specific language.\nIn particular\, this easily allows changi
ng the backend used in the end. `CFL` also provides the possibility to gen
erate the Newton linearization from a residual operator described in this
language.\n\nhttps://indico.sissa.it/event/23/session/1/contribution/10
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/1/contribution/10
END:VEVENT
BEGIN:VEVENT
SUMMARY:Accelerating augmented and deflated Krylov space methods for conve
ction-diffusion problems
DTSTART;VALUE=DATE-TIME:20180723T090000Z
DTEND;VALUE=DATE-TIME:20180723T092000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-13@cern.ch
DESCRIPTION:Speakers: PITTON\, Giuseppe (SISSA)\nIn this talk I will recal
l some basic notions of augmented and deflated Krylov space methods for th
e iterative solution of linear systems.\nThen I will discuss a few strateg
ies to apply these techniques to the solution of linear systems coming fro
m nonlinear scalar convection-diffusion equations.\nA simple back-of-the-e
nvelope analysis allows to argue that in some cases it may be convenient t
o exploit alternative recycling strategies based on the SVD selection of p
revious solutions\, and this conclusion is corroborated by numerical tests
based on Finite Element and Spectral Element discretisations.\n\nhttps://
indico.sissa.it/event/23/session/2/contribution/13
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/13
END:VEVENT
BEGIN:VEVENT
SUMMARY:An overview of some new features of deal.II 9.0
DTSTART;VALUE=DATE-TIME:20180727T070000Z
DTEND;VALUE=DATE-TIME:20180727T074500Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-12@cern.ch
DESCRIPTION:Speakers: Prof. HELTAI\, Luca (SISSA)\nVersion 9.0.0 of the `d
eal.II` library offers many new features and interfaces with several new e
xternal libraries. \n\nIn this talk I will provide an overview of some of
the new features that I have contributed in adding to the library\, includ
ing (in no particular order)\n\n - Improved support for curved geometries
\n - Non-standard quadrature rules\n - Improved support for user-defined r
un-time parameters\n - Support for GMSH\n - Support for Assimp\n - Support
for NanoFlann\n - Support for SUNDIALS\n\nhttps://indico.sissa.it/event/2
3/session/1/contribution/12
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/1/contribution/12
END:VEVENT
BEGIN:VEVENT
SUMMARY:Fluid-Structure Interaction: simulating the cavitation phenomenon
DTSTART;VALUE=DATE-TIME:20180724T090000Z
DTEND;VALUE=DATE-TIME:20180724T092000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-14@cern.ch
DESCRIPTION:Speakers: ABBONDANZA\, Dario (Sapienza University of Rome)\nIt
has been recently shown that a Diffuse Interface model is able to fully d
escribe the dynamics of multiphase flows. In particular\, the possibility
of deriving a set of equations addressing the dynamic of the entire domain
of simulation\, makes it ideal to describe phenomena such as vapor bubble
dynamics in a surrounding liquid. The model is able to capture surface te
nsion effects\, phase changes (condensation and evaporation) and even tran
sitions to supercritical conditions\, bubble collapse stages and the relat
ed shock-wave emissions. The diffuse interface model is intrinsically mult
iscale\, enabling the study of all the characteristic size of the bubbles
(from nm up to mm of radius) and the typical frequencies of evolution (of
order 1-100 MHz). Moreover\, the numerical simulations gives access to all
the fields\, anywhere in the domain\, in particular predicting the pressu
re and thermal loads on the nearby solid walls.\nA main application we are
interested in is the cavitation phenomenon (implosion of vapor bubbles ne
ar solid surfaces)\, in order to take advantage of the high mechanical str
esses (for biomedical applications\, e.g.) or prevent them (to reduce or p
revent structural damages). A full understanding of this complex phenomena
requires\, on one hand\, a full description of the fluid part\, and on th
e other hand\, a suitable modeling of the solid dynamics with particular a
ttention to plasticity and damage analysis. In this context\, the use of t
he deal.II library could help thanks to its ability in automatically refin
ing the mesh (i.e. in solving the smallest scales involved in the dynamics
)\, and thanks to the large implementation possibility\, already offered f
rom the structural point of view\, in terms of plasticity models.\n\nhttps
://indico.sissa.it/event/23/session/2/contribution/14
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/14
END:VEVENT
BEGIN:VEVENT
SUMMARY:Non matching grids in a distributed setting
DTSTART;VALUE=DATE-TIME:20180725T090000Z
DTEND;VALUE=DATE-TIME:20180725T092000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-17@cern.ch
DESCRIPTION:Speakers: Mr. ALZETTA\, Giovanni (SISSA)\nI am beginning this
talk with a quick recall of the general difficulties faced when using mult
iple grids on a distributed memory setting and some possible solutions.\nA
s a model problem I am presenting briefly the Lagrange-multiplier method o
f step-60\, where a harmonic equation with values prescribed at the bounda
ry and on a sub-domain is solved\, with a particular focus on the coupling
matrix used. After discussing how a partial solution making use of boundi
ng boxes was implemented in deal.II 9.0\, I am discussing its current limi
tations and intended developements.\n\nhttps://indico.sissa.it/event/23/se
ssion/2/contribution/17
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/17
END:VEVENT
BEGIN:VEVENT
SUMMARY:Model Order Reduction and Computational Homogenisation of Magnetor
heological Elastomers
DTSTART;VALUE=DATE-TIME:20180724T092000Z
DTEND;VALUE=DATE-TIME:20180724T094000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-16@cern.ch
DESCRIPTION:Speakers: Mr. BRANDS\, Benjamin (Friedrich-Alexander-Universit
ät Erlangen-Nürnberg)\nThe mechanical response of magnetorheological com
posites is highly affected by an applied magnetic field. Since a generally
valid constitutive law does not exist for such heterogeneous materials\,
multiscale techniques like computational homogenisation are commonly used
to approximate effective macroscopic properties. In our approach the macro
scopic quantities at a material point of a magnetorheological elastomer ar
e derived from the response of the underlying micro-structure\, where the
constitutive law is known\, using first-order homogenisation.\n\nThe compu
tational cost of this nested solution scheme known as the FE² method proh
ibits the simulation of complex macroscopic problems. To mitigate the comp
utational bottleneck the FE models on the microscale are replaced by reduc
ed-order models (ROMs). In projection-based ROM the governing equations ar
e projected onto the reduced basis\, which is an approximation of the solu
tion manifold of the parametrised partial differential equations (pPDE). T
he reduced basis is commonly constructed using previously computed solutio
ns of the pPDE\, e.g. by applying proper orthogonal decomposition or the r
educed basis method.\n\nWe will present our approach for the construction
and computation of the reduced-order models on the microscale. Through var
ious numerical examples the accuracy and time savings of the reduced model
s will be discussed.\n\nhttps://indico.sissa.it/event/23/session/2/contrib
ution/16
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/16
END:VEVENT
BEGIN:VEVENT
SUMMARY:Multilevel solver for discontinuous viscosity Stokes problem
DTSTART;VALUE=DATE-TIME:20180725T092000Z
DTEND;VALUE=DATE-TIME:20180725T094000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-19@cern.ch
DESCRIPTION:Speakers: WICHROWSKI\, Michal (Instytut Podstawowych Problemó
w Techniki)\nThe goal of my work is to develop robust scalable solver for
linear system arising from finite element discretization of Stokes proble
m with strongly variable viscosity. Such linear systems appear in fluid-st
ructure problems. \nSymmetric positive define problems with discontinuous
coefficient can be solved effectively by multigrid preconditioned Krylov
subspace methods. Here we extend this idea on saddle point problems by us
ing multilevel preconditioned GMRes. The multilevel method developed by Br
eass and Sarazin and further refined by Zulehner is applied. A multigrid a
lgorithm works on the saddle point problem by applying constrained smoothe
r. The relaxation allows the iterative procedure to remain in divergence f
ree subspace while smoothing out the error in both pressure and velocity.
Similarly to other multilevel methods\, the algorithm presented here does
not require explicitly storing the matrix and thus\, deal.II matrix-free
framework can be used.\n\nhttps://indico.sissa.it/event/23/session/2/contr
ibution/19
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/19
END:VEVENT
BEGIN:VEVENT
SUMMARY:Deal.II in cosmic ray propagation
DTSTART;VALUE=DATE-TIME:20180723T094000Z
DTEND;VALUE=DATE-TIME:20180723T100000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-18@cern.ch
DESCRIPTION:Speakers: WANG\, Jiaxin (SISSA)\nCosmic ray propagation in flu
id approach can be described as a PDE system\,\n\nwhich lives in 6D spatia
l-spectral space.\n\nIn our project\, we attempt to build up a solver for
such a system\, \n\nby combining independent triangulation in two domains
(spatial and spectral).\n\nIn this talk\, I will present why and how we bu
ild this framework with Kronecker\n\nproduct.\n\nhttps://indico.sissa.it/e
vent/23/session/2/contribution/18
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/18
END:VEVENT
BEGIN:VEVENT
SUMMARY:Integration and application of symbolic and automatic differentiat
ion frameworks in deal.II
DTSTART;VALUE=DATE-TIME:20180724T070000Z
DTEND;VALUE=DATE-TIME:20180724T074500Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-1@cern.ch
DESCRIPTION:Speakers: Dr. PELTERET\, Jean-Paul (Chair of Applied Mechanics
\, Friedrich-Alexander University Erlangen-Nuremberg)\nIn recent years the
re has been a notable growth in research among fields involving coupled me
dia\, and in multi-physics problems in general. As the complexity of the p
hysics describing the material loading and response increases\, so does th
e time cost and sophistication of the computational tools used to analyse
these problems both in the academic and industrial settings. Associated wi
th this are further challenges\, one example being that potential errors i
n implementation and subsequent numerical instabilities could be misinterp
reted as physically plausible material instabilities.\n\nTools that automa
te differentiation are one mechanism to tackle the challenges of constitut
ive law and finite element formulation prototyping\, development and valid
ation. In this work we discuss the implementation of frameworks to perform
assisted differentiation within the deal.II finite element library. We ou
tline and discuss some of the challenges surrounding the integration of tw
o algorithmic (automatic) and symbolic differentiation libraries within th
e pre-established tensor and finite element frameworks. Lastly\, we discus
s applications in the simulation of magnetorheological elastomers\; these
are smart\, field-responsive composite materials that are of increasing in
terest in numerous industries. We demonstrate and benchmark the current im
plementation in the context of two problems\, namely (i) a rate-dependent
magnetic field-response constitutive law and (ii) homogenisation of repres
entative volume element for a rate-independent magneto-rheological elastom
er.\n\nhttps://indico.sissa.it/event/23/session/1/contribution/1
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/1/contribution/1
END:VEVENT
BEGIN:VEVENT
SUMMARY:Welcome remarks. Past and future of deal.II
DTSTART;VALUE=DATE-TIME:20180723T070000Z
DTEND;VALUE=DATE-TIME:20180723T074500Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-0@cern.ch
DESCRIPTION:Speakers: Prof. BANGERTH\, Wolfgang (Colorado State University
)\ndeal.II is now 20 years old. I'll give a brief overview of how it start
ed\, how it has developed\, and where I see it going in the future. This i
ncludes both the technical side and\, in particular\, the social side: dea
l.II is not just a software project\, but also a community of developers a
nd users.\n\nhttps://indico.sissa.it/event/23/session/1/contribution/0
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/1/contribution/0
END:VEVENT
BEGIN:VEVENT
SUMMARY:Closing remarks
DTSTART;VALUE=DATE-TIME:20180727T103000Z
DTEND;VALUE=DATE-TIME:20180727T110000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-3@cern.ch
DESCRIPTION:Speakers: Prof. BANGERTH\, Wolfgang (Colorado State University
)\, HEISTER\, Timo (Clemson University)\, Prof. HELTAI\, Luca (SISSA)\nSom
e conclusions and some plans for future developments and opportunities. In
particular:\n\n - How to attract funds to `deal.II`\n - How to attract ne
w users\n - Open discussions on the best platforms for Forums/Discussions/
Mailinglists\n - Anything else\n\nhttps://indico.sissa.it/event/23/contrib
ution/3
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/contribution/3
END:VEVENT
BEGIN:VEVENT
SUMMARY:Supports for GPU in deal.II
DTSTART;VALUE=DATE-TIME:20180725T070000Z
DTEND;VALUE=DATE-TIME:20180725T074500Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-2@cern.ch
DESCRIPTION:Speakers: TURCKSIN\, Bruno (ORNL)\nGPUs are extensively used i
n supercomputers both in Europe with Piz Daint and in the United States wi
th Titan\, Sierra\, and Summit. The main reason is that they offer great f
loating points performance for a given power envelope. GPUs are also the d
e facto standard architectue for neural networks whose usage has been incr
easing quite rapidly in the last few years. It is therefore very advantage
ous if other applications can use the same infrastructure as the one used
for AI. However\, progamming on GPU is not the same as programming on CPU.
GPUs offer a lot more parallelism but GPU cores are a lot less powerful t
han CPU cores.\n\nIn this talk\, we will discuss how to use deal.II on GPU
s for matrix-based and matrix-free applications. For matrix-based applicat
ions\, we rely on cuSPARSE and cuSOLVER for operations on sparse matrices
and for some solvers. Matrix-free methods rely on the fact that many moder
n codes are using Krylov methods as their solver of choice. These solvers
do not require the matrix of the system but only the result of a sparse ma
trix-vector multiplication. This operation is memory bound both on CPU and
GPU. Matrix-free methods reduce drastically the memory requirement and th
us\, allow for a better utilization of the hardware. Moreover\, the extens
ive use of tensor products make these methods extremely interesting for hi
gh-order finite elements. High-order finite elements are attractive not on
ly because of the convergence properties but also because they introduce s
tructure even on unstructured meshes. Finally\, the lesser memory requirem
ent of matrix-free algorithms is very advantageous for GPUs since they do
not have a lot of memory per core. We will show some results and discuss o
ur path forward.\n\nhttps://indico.sissa.it/event/23/session/1/contributio
n/2
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/1/contribution/2
END:VEVENT
BEGIN:VEVENT
SUMMARY:Creative freedom for computational mesh generation in deal.II
DTSTART;VALUE=DATE-TIME:20180725T074500Z
DTEND;VALUE=DATE-TIME:20180725T083000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-5@cern.ch
DESCRIPTION:Speakers: Mr. GIULIANI\, Nicola (SISSA)\nMany modern 3D modell
ing tools offer the possibility of creating complex geometries in a very s
traightforward way (what is usually called creative freedom). Such softwar
es are extensively used in computer graphics (video-gaming\, animation stu
dios) and they are able to produce smooth surfaces exploiting the subdivis
ion surface modelling which was derived by Edwin Catmull and Jim Clark in
1978 as a generalisation of bi-cubic uniform B-spline surfaces to arbitrar
y topology. We believe that a bridge between such computer graphics techni
ques and scientific Finite Element computations can open new perspectives
for the deal.II users community.\nWe present the possibility to use such t
ools to generate a complex geometry to be integrated as a computational me
sh inside the deal.II library. \n\nTo demonstrate the possibilities of our
approach we use Blender\, an Open Source python based software which has
emerged as one of the most successful 3D modelling tool. We believe that i
ts success is due both to its outstanding possibilities and spreading\, in
particular we stress that Blender \n\n- is Open Source software\,\n- allo
ws for straightforward geometry creation using Graphic User Interface\,\n-
is highly scriptable using a Python 3 based Text User Interface\,\n- has
a large user community (tutorials\, manuals\, youtube channel\, wikipedia)
.\n\n\nWe use the Open Asset Import Library (ASSIMP) to convert the standa
rd Blender format into a deal.II geometry through the interface that is al
ready inside the library. These passages allow deal.II users to generate c
omplex three dimensional surfaces in a very controllable way. \n\nWe prove
the capabilities of our approach in different framework varying from Boun
dary Element Method for Stokes flows\, to the discretisation of surface op
erators on non trivial geometries.\n\nhttps://indico.sissa.it/event/23/ses
sion/1/contribution/5
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/1/contribution/5
END:VEVENT
BEGIN:VEVENT
SUMMARY:Numerical (micro)aquarium using the deal.II library
DTSTART;VALUE=DATE-TIME:20180724T094000Z
DTEND;VALUE=DATE-TIME:20180724T095000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-4@cern.ch
DESCRIPTION:Speakers: Dr. GIULIANI\, Nicola (SISSA)\nDuring the last 8 yea
rs the mathLab laboratory at SISSA has analysed the swimming mechanisms of
micro-organisms (unicellular beings as protists\, algae or bacteria) both
from a theoretical and an experimental point view. In this work we presen
t a way to derive an accurate and reliable numerical setting for the resol
ution of general swimming mechanism. Such a framework can be seen as a num
erical microscope.\n\nA general swimming motion (both for micro than macro
organism) can be seen as a Fluid Structure Interaction problem. \nGiven t
he characteristic length-scale of the problem (100 microns at most) the fl
uid part of the problem is well described by the Stokes equations (in part
icular we use the Boundary Integral Equations for the Stokes system). For
what concerns the structural part of the system the typical swimming mecha
nisms (flagellar movements\, body deformations) involve large domain defor
mations which pose severe mesh quality problems to finite element or finit
e volume solvers\, and make Boundary Element Method an ideal discretisatio
n strategy for these FSI problems\, requiring only the discretisation of t
he domain boundaries.\n\nWe present a series of tools that allows the reso
lution of swimming problem starting from the analysis of an experimental v
ideo. In particular we isolate the shape changes of the swimmer and we use
this informations as input data to recover the rigid movements of the ove
rall organism. We use an OpenSource 3D modelling tool (Blender) to generat
e the computational mesh and we use an efficient MPI parallel solver (base
d on deal.II) for Stokes problems exploiting a collocation Boundary Elemen
t Method. The BEMStokes library\, released under LGPL license\, solve both
2D and 3D problem and it is rigorously tested using continuous integratio
n over more than 120 different test cases. \n\nThe resulting softwares con
stitutes what we call a "numerical aquarium" since it allows for an effici
ent and reliable study of complex swimming organisms complementing the ana
lysis that can be obtained through a real microscope.\n\nhttps://indico.si
ssa.it/event/23/session/2/contribution/4
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/4
END:VEVENT
BEGIN:VEVENT
SUMMARY:Algorithms and data structures for matrix-free finite element oper
ators with MPI-parallel sparse multi-vectors
DTSTART;VALUE=DATE-TIME:20180726T070000Z
DTEND;VALUE=DATE-TIME:20180726T074500Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-7@cern.ch
DESCRIPTION:Speakers: DAVYDOV\, Denis (Friedrich-Alexander University Erla
ngen-Nuremberg)\nTraditional solution approaches for problems in quantum m
echanics scale as $\\mathcal O(N^3)$\, where $N$ is the number of electron
s. Various methods have been proposed to address this issue and obtain lin
ear scaling $\\mathcal O(N)$. One promising formulation is the direct mini
mization of energy. Such methods take advantage of physical localization o
f the solution\, namely that the solution can be sought in terms of non-or
thogonal orbitals with local support. This is often called the near-sighte
dness principle of matter.\n\nIn this talk we present numerically efficien
t implementation of sparse parallel vectors within the deal.II open-source
finite element library suitable for matrix-free operator evaluation. Base
d on the a-priori chosen support for each vector\, we develop algorithms a
nd data structures to perform (i) matrix-free sparse matrix multi-vector
products (SpMM) (ii) projection of an operator onto a sparse sub-space (in
ner products) (iii) post-multiplication with a matrix. Strong and weak sca
ling results are reported for a typical benchmark problem using quadratic
and quartic finite element bases.\n\nhttps://indico.sissa.it/event/23/sess
ion/1/contribution/7
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/1/contribution/7
END:VEVENT
BEGIN:VEVENT
SUMMARY:Numerical Approximation of the Integral Fractional Laplacian
DTSTART;VALUE=DATE-TIME:20180726T074500Z
DTEND;VALUE=DATE-TIME:20180726T083000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-6@cern.ch
DESCRIPTION:Speakers: Dr. LEI\, Wenyu (Texas A&M University)\nIn this talk
\, we propose a new nonconforming finite element algorithm to approximate
a linear problem involving the integral fractional Laplacian. The method i
s based on the Dunford-Taylor integral representation of the bilinear form
corresponding to the variational formulation. We obtain the approximated
solution using the deal.II finite element library with a non-uniform mesh
setting. Finally\, as an application\, we consider an obstacle problem inv
olving the integral fractional Laplacian and several numerical simulations
are also provided.\n\nhttps://indico.sissa.it/event/23/session/1/contribu
tion/6
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/1/contribution/6
END:VEVENT
BEGIN:VEVENT
SUMMARY:Fast Tensor Product Schwarz Smoothers
DTSTART;VALUE=DATE-TIME:20180725T094000Z
DTEND;VALUE=DATE-TIME:20180725T100000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-9@cern.ch
DESCRIPTION:Speakers: WITTE\, Julius (Heidelberg University)\nThe goal of
the ExaDG project is to provide efficient finite element solvers relying o
n `deal.II`'s `MatrixFree` framework based on higher order DG methods. How
ever\, the matrix-free operators restrict the choice of preconditioners. A
powerful tool is offered by a geometric multigrid algorithm having a favo
rable linear complexity. The quality of multigrid preconditioners is cruci
ally dependent on the choice of the smoother. Overlapping Schwarz smoother
s using local spaces supported on vertex patches were introduced by Arnold
\, Falk\, and Winther and provide robust convergence in particular for div
ergence-constrained problems and almost incompressible elasticity. In clas
sic finite element codes\, the local stiffness matrices are extracted from
the global system matrix. Therefore\, such smoothers are expensive to com
pute and store\, especially for higher order DG discretizations. Exploitin
g again the tensor product finite elements on the subproblem level offers
a remedy to both and fits well into a matrix-free framework. Thus\, we app
roximate the global bilinear form by separable problems in the smoother.\n
\nFirst\, we show how to reduce the complexity of inverting and storing su
bspace problems by means of the fast diagonalization method. Second\, a ge
neral tensor product setting is presented to illustrate the sum-factorizat
ion approaches used in finite element related matrix-vector products resul
ting from the global forward as well as local inverse problems. Finally\,
the efficiency of our proposed fast diagonalized Schwarz smoothers is unde
rlined by numerical experiments and we close with a brief overview of futu
re plans.\n\nhttps://indico.sissa.it/event/23/session/2/contribution/9
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/9
END:VEVENT
BEGIN:VEVENT
SUMMARY:Present and Future of adaptive\, parallel\, geometric multigrid in
deal.II
DTSTART;VALUE=DATE-TIME:20180723T074500Z
DTEND;VALUE=DATE-TIME:20180723T083000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-8@cern.ch
DESCRIPTION:Speakers: HEISTER\, Timo (Clemson University)\nMultigrid metho
ds are the only known approach to efficiently solve large\nlinear systems
for solving PDEs by having runtime cost proportional to the\nnumber of unk
nowns. Algebraic multigrid methods construct a hierarchy of\nproblems from
the system matrix\, while geometric multigrid methods use the\nmesh hiera
rchy and system matrices from each level of this hierarchy.\n\nWhile algeb
raic multigrid methods have been successful in the past and are\nstill the
method of choice for large problems in deal.II\, they have several\ndisad
vantages over geometric multigrid methods when implemented in a\nmatrix-fr
ee version: First\, scaling to more than a few thousand cores is\nlimited
due to the large setup cost. Second\, the low arithmetic intensity does\nn
ot lend itself to modern architectures. Third\, implementation of\nmultith
reading is difficult. Fourth\, the matrices consume large amounts of\nsyst
em memory.\n\nHere we present the current status of adaptive\, parallel\,
geometric multigrid\nhow it is implemented in deal.II. Several numerical t
ests demonstrate the\napproach. Finally\, we will close with future plans.
\n\nhttps://indico.sissa.it/event/23/session/1/contribution/8
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/1/contribution/8
END:VEVENT
BEGIN:VEVENT
SUMMARY:deal.II interface with CAD date structures: description\, results
and open issues
DTSTART;VALUE=DATE-TIME:20180726T090000Z
DTEND;VALUE=DATE-TIME:20180726T092000Z
DTSTAMP;VALUE=DATE-TIME:20190120T162225Z
UID:indico-contribution-23-21@cern.ch
DESCRIPTION:Speakers: Dr. MOLA\, Andrea (SISSA)\nTo exploit the automated
mesh generation and adaptive refinement capabilities offered by deal.II\,
the user must provide its program with a suitable description of the domai
n geometry. In particular\, it is fundamental that the geometries of bound
ary and internal surfaces are available to the triangulation so that the n
odes generated across each mesh refinement are placed onto such surfaces\,
in a way that preserves mesh quality. This task is currently carried out
by the manifold classes. A wide variety of manifold descriptors for the mo
st common analytical geometries is available in the library. Along with su
ch useful instruments\, a set of manifold descriptors which import and int
errogate CAD files has been implemented to treat also arbitrary geometries
. All such manifolds are based on wrappers of functions contained in the O
penCASCADE library\, that are included in the corresponding deal.II namesp
ace. This contribution will discuss such manifold descriptors and their wo
rking principles\, mentioning relevant details of their implementation and
providing examples and results of their applications. The discussion will
then move to a series of open issues which currently limit the OpenCASCAD
E manifolds computational efficiency.\n\nhttps://indico.sissa.it/event/23/
session/2/contribution/21
LOCATION:SISSA Main Building Meeting Room -- VII Floor
URL:https://indico.sissa.it/event/23/session/2/contribution/21
END:VEVENT
END:VCALENDAR