Subject: NA Digest, V. 16, # 14

NA Digest Tuesday, April 04, 2016 Volume 16 : Issue 14

Today's Editor:

  Daniel M. Dunlavy
  Sandia National Labs

Today's Topics:

Tensorlab 3.0 Release
European Trilinos User Group Meeting, Germany, Apr 2016
Structured Matrix Days, France, May 2016
Hybrid Simulation Methods in Fluid Dynamics, Germany, Jun 2016
Parameter Estimation for Biological Models, USA, Jul 2016
Computer Aspects of Numerical Algorithms (CANA'16), Poland, Sep 2016
Deadline extension, Auto-Tuning for Multicore and GPU, France, Sep 2016
Numerical Linear Algebra and Optimization, UK, Sep 2016
Postdoc Position, M2NeT Lab, WLU, Waterloo, Canada
Postdoc Position, Parallel Multigrid Methods for DG, Lund Univ
Postdoc/Researcher Position, Dynamism-aware HPC, NTNU, Norway
PhD/Doctorate Positions, HPC: An Exascale Hyperbolic PDE Engine
Contents, Numerical Algorithms, 71 (4)

Subscribe, unsubscribe, change address, or for na-digest archives:

Submissions for NA Digest:


From: Lieven De Lathauwer
Date: March 30, 2016
Subject: Tensorlab 3.0 Release

It is our pleasure to announce the release of Tensorlab 3.0.
Tensorlab is a Matlab toolbox for tensor computations, structured data
fusion and complex optimization.

In Tensorlab 3.0, dedicated algorithms for the decomposition in
multilinear rank-(Lr,Lr,1) terms have been added. The new release
introduces a tensorization framework and offers increased support for
handling large-scale and structured datasets. The performance,
user-friendliness and capabilities of the domain specific language for
structured data fusion have been enhanced, and a new tool has been
developed for the visualization of tensors of arbitrary order.

An extensive new online manual has been developed. A collection of
demos illustrates how Tensorlab can be used in various
applications. Code, manual, demos and detailed release notes are
available from the renewed website

The toolbox is free to use for non-commercial research. We hope that
you will find it useful in your research. We welcome all suggestions
and feedback.


From: Michael Heroux
Date: March 31, 2016
Subject: European Trilinos User Group Meeting, Germany, Apr 2016

The 5th European Trilinos User Group Meeting (EuroTUG2016) will take
place April 18-20 2016 in Garching near Munich at the Leibniz
Rechenzentrum (LRZ), jointly hosted this year by LRZ and TU Munich.

EuroTUG is an annual meeting that brings together researchers,
developers and users of the open source scientific software collection
'Trilinos' ( from Sandia National
Laboratories, USA. The meeting is a platform for discussions and
information on latest trends in HPC scientific software, massively
parallel computing, efficient algorithms and HPC enabling
technologies, with an emphasis on Trilinos usage.

EuroTUG will supply an overview of the Trilinos software suit as well
as beginners and advanced tutorials. One of this year-s focuses is on
usage of Trilinos-Kokkos, a programming model for performance portable
applications targeting all major platforms including Nvidia, Xeon Phi
and manycore complex hardware node architectures. Details can be
found at:

If you are interested in presenting your Trilinos-related work, please
indicate so when registering or by email to


From: Paola Boito
Date: March 31, 2016
Subject: Structured Matrix Days, France, May 2016

Structured Matrix Days
XLIM, Universite de Limoges
May 9 - 10, 2016

The Department of Mathematics and Computer Science at XLIM,
Universite de Limoges, organizes a two-day workshop on
structured matrices and their applications.

Confirmed invited speakers:
- Francesca Arrigo (Universita dell'Insubria),
- Robert Luce (EPFL),
- Clement Pernet (Universite Grenoble Alpes-LIP),
- Leonardo Robol (KU Leuven).

Deadline for registration and abstract submission:
April 25, 2016. For additional information, please
write to .


From: Philipp Neumann
Date: April 04, 2016
Subject: Hybrid Simulation Methods in Fluid Dynamics, Germany, Jun 2016

A workshop "Hybrid Simulation Methods in Fluid Dynamics: Models,
Software, and Applications" takes place on Jun 20-21 2016 at the
Technical University of Munich, Germany, Institute for Advanced Study.
The workshop will focus on the interplay of the disciplines
physics/math (modeling), computer science (HPC/software design), and
engineering (applications) to tackle current challenges in hybrid (in
particular molecular-continuum) simulation methods.

Invited speakers:
- Rafael Delgado-Buscalioni, Universidad Autonoma de Madrid
- Dimitris Drikakis, University of Strathclyde
- George Karniadakis, Brown University
- Kurt Kremer, Max Planck Institute for Polymer Research
- Duncan Lockerby, University of Warwick
- Nikolaus Adams, Technical University of Munich

The number of participants (ca. 25) is limited. There is no
registration fee, and up to two nights of accommodation can also be
covered. Please register for the workshop before April 22 2016;
details on the workshop and registration can be found at:


From: Andrea Arnold
Date: March 30, 2016
Subject: Parameter Estimation for Biological Models, USA, Jul 2016

Tutorial Workshop on Parameter Estimation for Biological Models
North Carolina State University, Raleigh, NC
July 28th-31st, 2016

The NC State Research Training Group on Parameter Estimation for
Mechanistic Biological Models (, funded by
the National Science Foundation (NSF), will run a tutorial workshop on
parameter estimation for biological models between July 28th and July
31st, 2016.

The primary target audience for the workshop is graduate students
interested in learning basic techniques associated with modern methods
of identifiability theory, parameter estimation, and uncertainty
quantification in models arising in biology. We also welcome
applications from advanced undergraduates, postdocs and others who may
be interested in these topics. Individuals wishing to participate
should contact Mette Olufsen ( Full consideration
will be given to applications received by May 1st, although
applications received after that date will be considered if space
permits. Financial support is available for a limited number of US
students and postdocs. If you wish to request financial support,
please attach a copy of your CV and ask your advisor to email a letter
of recommendation to us.

Further details about the workshop will appear at


From: Przemyslaw Stpiczynski
Date: April 04, 2016
Subject: Computer Aspects of Numerical Algorithms (CANA'16), Poland, Sep 2016

9th Computer Aspects of Numerical Algorithms (CANA'16)
Gdansk, Poland, 11-14 September, 2016

Numerical algorithms are widely used by scientists engaged in various
areas. There is a special need of highly efficient and easy-to-use
scalable tools for solving large scale problems. The workshop is
devoted to numerical algorithms with the particular attention to the
latest scientific trends in this area and problems related to
implementation of libraries of efficient numerical algorithms. The
goal of the workshop is meeting of researchers from various institutes
and exchanging of their experience, and integration of scientific

Topics (including but not limited to): Parallel numerical algorithms;
Novel data formats for dense and sparse matrices; Libraries for
numerical computations; Numerical algorithms testing and benchmarking;
Analysis of rounding errors of numerical algorithms; Languages, tools
and environments for programming numerical algorithms; Numerical
algorithms on GPUs; Paradigms of programming numerical algorithms;
Contemporary computer architectures; Heterogeneous numerical
algorithms; Applications of numerical algorithms in science and

Paper submission:

Important Dates:
Paper submission: April 18 2016 12:00:00 pm HST


From: Satoshi Ohshima
Date: April 02, 2016
Subject: Deadline extension, Auto-Tuning for Multicore and GPU, France, Sep 2016

5th Auto-Tuning for Multicore and GPU (ATMG 2016)

September 21-23, 2016 at Lyon Congress Center, Lyon, France.

Topics of Interest. Papers are solicited in following areas of
auto-tuning technologies for multicore and GPU/MIC computing, but not
limited to:
- Optimized Algorithms for Numerical Libraries
- Automatic Code Generation and Empirical Compilation Hybrid
Programming for Threads and Processes
- Communication Optimization
- Mixed Precision and Accuracy Assurance for Numerical Computing
- Power Consumption Optimization
- Fault Tolerance

Paper submission due : April 30, 2016
Author notification : June 24, 2016


From: Pamela Bye
Date: April 01, 2016
Subject: Numerical Linear Algebra and Optimization, UK, Sep 2016

Fifth IMA Conference on Numerical Linear Algebra and Optimization
University of Birmingham, UK
7 - 9 September 2016
Conference webpage:

Conference topics include any subject that could be of interest to
both communities, such as: Direct and iterative methods for large
sparse linear systems; Eigenvalue computation and optimization;
Large-scale nonlinear and semidefinite programming; Effect of
round-off errors, stopping criteria, embedded iterative procedures;
Optimization issues for matrix polynomials; Fast matrix computations;
Compressed/sparse sensing; PDE-constrained optimization; Distributed
computing and optimization; Applications and real time optimization.

Mini-symposium proposals and contributed talks are invited on all
aspects of numerical linear algebra and optimization. Mini-symposium
proposals should be submitted to by 31 March
2016. A mini-symposium is limited to at most two sessions on a single
topic (maximum eight speakers). Organisers will be advised of
acceptance by 11 April 2016. Contributed talks and mini-symposia talks
will be accepted on the basis of a one page extended abstract which
should be submitted by 30 April 2016 by e-mail to Authors will be advised of acceptance by 31
May 2016. A book of abstracts will be made available to delegates at
the conference.

For further information or to register your interest, please contact
Lizzi Lake, Conference Officer, E-mail:


From: Roderick Melnik
Date: April 01, 2016
Subject: Postdoc Position, M2NeT Lab, WLU, Waterloo, Canada

Applications are invited for a Postdoctoral Position in Modeling for
Complex Systems, with focus on nanoscale systems and biophysical
applications. The position is based at the M2NeT Lab which is part of
the MS2Discovery Interdisciplinary Research Institute, located at
Wilfrid Laurier University, Waterloo, Canada. Further information
about the position and how to apply can be found at the following

It is expected that that full consideration will be given to all
applications received by May 30, 2016, but applications will be
reviewed as they are received.

The position is available from the Fall of 2016.


From: Philipp Birken
Date: April 01, 2016
Subject: Postdoc Position, Parallel Multigrid Methods for DG, Lund Univ

We have an opening at the numerical analysis group at Lund University
for a PostDoc position for one year. Candidates should contact me for
questions. The deadline is April 20th and applications are done online
via the complete description of the announcement at:

The research will be focused on the analysis and design of parallel
multigrid methods in the context of Discontinuous Galerkin
discretizations for unsteady compressible flows. The aim is to design
scalable, fast and efficient methods for the next generation of
solvers for Large Eddy Simulations (LES). This will be in
collaboration with other groups in Sweden, Germany and the US.

About 20% of the working hours are for helping several research groups
at the Faculty of Science and the Faculty of Engineering on code
optimization of their high performance computing applications.

Teaching duties in numerical analysis of up to 20 % can be included in
the job assignment.

The candidate must have a PhD in Numerical Analysis or similar, not
older than three years. Good knowledge of high performance computing
is required. Good knowledge of high order discretizations for Partial
Differential Equations (PDEs) and Discontinuous Galerkin methods
specifically is desired. Knowledge of multigrid methods, numerical
linear algebra and computational fluid dynamics is desired.

Good programming skills in Fortran or C++ and demonstrable experience
with parallel implementation of numerical methods for PDEs are

Good collaboration skills and an interest in interdisciplinary work
are required.


From: Magnus Jahre
Date: April 04, 2016
Subject: Postdoc/Researcher Position, Dynamism-aware HPC, NTNU, Norway

A postdoc/researcher position is available within the "Run-time
Exploitation of Application Dynamism for Energy-efficient eXascale
computing (READEX)" Horizon 2020 EU project at NTNU, Trondheim, Norway
with application deadline April 24 2016.

The objective of the READEX project is to develop a design methodology
and tool-suite that enables exploiting dynamic application behaviour
in HPC systems for improved energy efficiency. The project will build
on, combine and extend research results from system scenario based
design in the embedded domain and HPC auto-tuning technology.

The postdoctoral fellow/researcher will mainly contribute to the following
tasks in collaboration with other researchers at NTNU:
(a) Identify dynamic behaviour in HPC applications
(b) Determine the performance and energy impact of changing platform
settings (e.g. voltages, clock frequencies, communication
algorithms, etc.) for each application phase
(c) Develop design time techniques that detect dynamic behaviour and
determine near-optimal platform settings for each application
(d) Develop low-overhead runtime techniques that select platform
settings for each phase based on information from the design time
analysis and additional dynamic measurements.

The complete job announcement can be found at:


From: Michael Bader
Date: March 31, 2016
Subject: PhD/Doctorate Positions, HPC: An Exascale Hyperbolic PDE Engine

PhD/Doctorate Positions in High Performance Computing:
An Exascale Hyperbolic PDE Engine -
at Technical University of Munich and at Durham University

ExaHyPE is a Horizon 2020 project (2015-2019) to develop a novel
simulation engine for hyperbolic PDE problems at exascale. We
especially tackle grand-challenge simulations in Seismology and
Astrophysics. The underlying simulation engine will be based on
high-order discontinuous Galerkin discretization on dynamically
adaptive Cartesian grids. To strengthen our team, we're looking for a
doctoral researcher who ideally should already have a strong
background on HPC topics such as (but not restricted to) MPI
programming, load balancing or performance optimization for multi- and
manycore CPUs. Interest in the application areas is
advantageous. However, we see the ability and passion to integrate and
learn new techniques as your most important qualification.

for full details.

Applications will be considered until the positions are filled. To
ensure a timely start at Oct 1, 2016, we appreciate applications
before May 15. Please send your application as a single PDF document

Michael Bader, Technische Universität München,
Tobias Weinzierl, Durham University,


From: Claude Brezinski
Date: April 04, 2016
Subject: Contents, Numerical Algorithms, 71 (4)

Table of contents for Numerical Algorithms, Volume 71, Issue 4.

Cascadic multigrid methods combined with sixth order compact scheme
for poisson equation, Li Ming, Li Chen-liang, Cui Xiang-zhao, Zhao

On a nonlinear mean and its application to image compression using
multiresolution schemes, S. Amat, J. Liandrat, J. Ruiz, J. C. Trillo

Improved PHSS iterative methods for solving saddle point problems, Ke
Wang, Jingjing Di, Don Liu

An optimal fourth-order family of methods for multiple roots and its
dynamics, Ramandeep Behl, Alicia Cordero, Sandile S. Motsa, Juan R.
Torregrosa, Vinay Kanwar

Global solutions of the biconfluent Heun equation, E. M. Ferreira,
J. Sesma

Improved convergence analysis for Newton-like methods, Ángel Alberto
Magrenan, Ioannis K. Argyros

Using generating functions to convert an implicit (3,3) finite
difference method to an explicit form on diffusion equation with
different boundary conditions, Saeed Kazem, Edmund Chadwick, Ali
Hatam, Mehdi Dehghan

On variable step Hermite-Birkhoff solvers combining multistep and
4-stage DIRK methods for stiff ODEs, Truong Nguyen-Ba

A deterministic sparse FFT algorithm for vectors with small support,
Gerlind Plonka, Katrin Wannenwetsch

Modify Levinson algorithm for symmetric positive definite Toeplitz
system, Nasser Akhoundi, Iman Alimirzaei

A generalized forward-backward splitting method for solving quasi
inclusion problems in Banach spaces, Prasit Cholamjiak

Semilocal convergence of an eighth-order method in Banach spaces and
its computational efficiency, J. P. Jaiswal

End of Digest
if (! preg_match( "/\d+/", $volume ) && preg_match( "/\d+/", $issue ) ) { exit( 'Invalid parameters' ); }