SIAM banner

SIAM Activity Group on Linear Algebra

SIAG LA digest

Apr 4, 2016 • Jennifer Pestana

Marvin Marcus passed away on February 20, 2016. Marvin will be remembered not only for his many contributions to matrix theory and multilinear algebra but also for his support and mentorship of others, particularly his graduate students. Among his many achievements, Marvin co-founded the journal Linear and Multilinear Algebra, and a special issue will be devoted to his memory as detailed below.

CONTENTS


Special issue of Linear and Multilinear Algebra in memory of Marvin Marcus

Dear colleagues,

The linear and multilinear algebra community is sad to lose Marvin Marcus, who had make substantial contributions on various topics on the subject, developed human resources by training and attracting many researchers to the research area, and co-founded the journal Linear and Multilinear Algebra. A special issue of Linear and Multilinear Algebra will be devoted to the memory of Marvin Marcus. Colleagues are encouraged to submit paper to the special issue by September 30, 2016. The special editors are:

Shmuel Friedland, University of Illinois - Chicago, [email protected]
Thomas Pate, Auburn University, [email protected]
Yiu-Tung Poon, Iowa State University, [email protected]

Submissions should be done through the website: https://mc.manuscriptcentral.com/glma

Steve Kirkland and Chi-Kwong Li
Editors, Linear and Multilinear Algebra


SISC Copper Mountain Special Section

The Fourteenth Copper Mountain Conference on Iterative Methods was held on March 20-25, 2016. The meeting attracted more than 200 participants from around the world, and there were 173 talks covering a broad spectrum of topics related to iterative solvers (http://grandmaster.colorado.edu/~copper/2016/).

As in previous even-numbered years, the SIAM Journal on Scientific Computing (SISC) is planning a Special Section, dedicated to recent progress in iterative methods.

Submissions are encouraged in all aspects of iterative methods including the following areas:

  • Iterative Linear Algebraic Data Mining Techniques
  • Optimization of Complex Problems/Systems (PDE-constrained optimization, optimization under uncertainty, and simulation-based optimization)
  • Inverse problems, regularization
  • Nonlinear solution methods, nonlinear least squares
  • Eigenvalue and singular value methods and applications
  • Iterative solvers/linear algebra on high concurrency node architectures
  • Uncertainty quantification/PDEs with random data
  • Hybrid direct-iterative solvers
  • Solvers for indefinite systems
  • Surrogate modeling/model reduction
  • Low-rank solvers for systems with Kronecker structure
  • Robust and scalable iterative solution of coupled multi-physics problems: electromagnetics/fluids
  • Iterative methods in energy applications
  • Iterative methods in imaging
  • Quantum chemistry/physics
  • Iterative solvers in climate applications

The window for submission is April 1 - June 30, 2016. Attendees and participants of the conference as well as the general community are invited to submit papers. Papers will be subject to review by a guest Editorial Board.

For submission instructions, visit

http://www.siam.org/journals/sisc/sisc_special.php

or

http://grandmaster.colorado.edu/~copper/2016/sisc.php

Michele Benzi Guest Editor-in-Charge, Copper Mountain Special Section


9th International Workshop on Parallel Matrix Algorithms and Applications (PMAA16)

IMPORTANT DATES

  • April 15, 2016 - Acceptance notification
  • May 13, 2016 - End of early registration
  • June 17, 2016 - Online registration closing
  • July 6-8, 2016 - Conference

CONFERENCE OVERVIEW

The 9th International Workshop on Parallel Matrix Algorithms and Applications (PMAA16) will be held in Bordeaux (France) and will run from July 6 to 8, 2016.

The workshop aims at being a forum for an exchange of ideas, insights and experiences in different areas of parallel computing (Multicore, Manycores and GPU) and applications in which matrix algorithms are employed. The PMAA16 workshop will bring together experts and researchers from diverse disciplines with a common interest in matrix computation. The purpose of the workshop is to discuss recent developments in various aspects of parallel matrix algorithms and to bring together mathematicians, computational scientists and engineers.

The workshop will focus on:

  • Parallel methods for solving large sparse or dense linear systems of equations.
  • Parallel eigenvalue and singular value computations.
  • Automatic tuning and performance modeling.
  • Novel architectural paradigms (e.g. accelerator and multicore) and their use in matrix computations and applications.
  • Large scale scientific applications from diverse fields which have an emphasis on parallel matrix computation.

We invite you to submit scientific contributions in all fields of numerical analysis, scientific and industrial applications, software development, and high performance computing. From the submissions received, the Organizing Committee will select papers to be presented as oral contributions. We hope that this opportunity to present and discuss science will interest you, and we look forward to seeing you in Bordeaux! Papers presented at the PMAA16 and submissions on closely related topics will be considered for publication in a special issue of of Parallel Computing after a peer review process. They must be submitted electronically via Elsevier’s online submission system at http://ees.elsevier.com/parco/.

###COMMITTEE

Co-chairs: E. Agullo (Inria, France), P. Arbenz (ETH Zurich, Switzerland), L. Giraud (Inria, France), O. Schenk (USI, Switzerland).

Steering committe: P. Arbenz (ETH Zurich, Switzerland), E. Kontoghiorghes (U London, UK) , Y. Saad (U Minnesota, USA), A. Sameh (Purdue U, USA), O. Schenk (USI, Switzerland).

Local organizing committee: E. Agullo, V. Cabet, O. Coulaud, A. Esnard, M. Faverge, L. Giraud, L. Grimaldi, C. Plumejeau, P. Ramet, J. Roman, S. Valerius.

International program committee: P. D’Ambra (ICAR, Italy), H. Avron (Tel Aviv U, Israel), M. Bader (TU Munchen, Germany), W. Bangerth (TAMU, USA), A. Basermann (BLR, Germany), C. Bekas (IBM Zurich, Switzerland), P. Bientinesi (Aachen U,Germany), R. Blaheta (TU Ostrava, Czech Republic), M. Bollhoefer (TU Brunswick, Germany), E. Boman (SNL, USA), G. Bosilca (UTK, USA), A. Buttari (CNRS, France), X.-C. Cai (UCB, USA), H. Calandra (Total, USA), O. Coulaud (Inria, France), E. Chow (Georgia Tech, USA), E. Darve (Stanford U, USA), Z. Drmac (U Zagreb, Croatia), N. Emad (UVSQ, France), M. Faverge (Bordeaux INP, France), E. Gallopoulos (U Patras, Greece), D. Goeddeke (TU Dortmund, Germany), J. Gondahmed samehzio (U Edinburgh, UK), D. Gordon (U Haifa , Israel), L. Grigori (Inria, France), A. Guermouche (U Bordeaux, France), I. Gutheil (JSC, Germany), G. Houzeaux (BSC, Spain), T. Huckle (TU Munchen, Germany), M. Knepley (RICE U, USA), R. Krause (USI, Switzerland), J. Langou (UCD, USA), H. Ltaief (KAUST, Saudi Arabia), K. Meerbergen (KU Leuven, Belgium), Y. U. Meier (LLNL, USA), E. Ng (LBNL, USA), G. Oksa (Slovak Academy of Sciences, Slovak Republic), S. Petiton (U. Lille, France), P. Ramet (U. Bordeaux, France), J. Roman (Inria, France), J. E. Roman (UPV, Spain), C. Petra (ANL, USA), T. Sakurai (U Tsukuba, Japan), D. B. Szyld (Temple U, USA), D. Trystram (Grenoble INP, France), M. Tuma (Praha U, Czech Republic), B. Uçar (CNRS, France), M. Vajtersic (Salzburg U, Austria), W. Vanroose (U Antwerp, Belgium), P. Vasconcelos (Porto U, Portugal), R. Wyrzykowski (TU Czestochowa, Poland).

###INVITED SPEAKERS

  • Cleeve Ashcraft (LSTC, Livermore)
  • Costas Bekas (IBM, Switzerland)
  • Thierry Deutsch (CEA, France)
  • Jack Dongarra (University of Tennessee, USA, Oak Ridge National Laboratory, USA, and University of Manchester, UK)

With best regards,

Emmanuel Agullo


CSCS-USI Summer School on Effective High Performance Computing

The Swiss National Supercomputing Centre (http://www.cscs.ch) and the Institute of Computational Science (http://www.ics.usi.ch) at Università della Svizzera italiana are pleased to announce the opening of the selection process for participation in the annual Summer School. The Summer School will be held from July 17 through July 28, 2016, in Hotel Serpiano (http://www.serpiano.ch), located in the Italian area of Switzerland. The hotel is positioned on Mount San Giorgio, UNESCO World Heritage site since 2003, wherefrom participants can enjoy an incredible view of Ceresio lake and its surroundings.

The Summer School will focus on the effective exploitation of High Performance Computing (HPC) systems, more details on the agenda can be found on our webpage: http://goo.gl/VYnFE8

Please see the flyer in attachment and the youtube video from the 2015 Summer School: https://goo.gl/UDwmST

###REQUIREMENTS

The Summer School is dedicated to Master and Ph.D. students, academic staff and professionals. Applicants have to match the following requirements:

  • Good knowledge of at least one of the following programming languages: C++, C, or Fortran 90
  • Familiarity with the UNIX/Linux OS and the basic programming environment tools (terminals, editors, compilers)
  • Require HPC skills for their work or research.

Only applicants who satisfy all requirements will be considered. Participants will need to bring their own laptop for practical sessions.

###CONTENT The school will focus on the effective exploitation of High Performance Computing (HPC) systems. Current and coming generations of HPC systems are increasing in size and complexity. Multi-core processors, general purpose graphic accelerators (GPUs), many-cores accelerators (MIC), ultra-fast parallel storage devices, and high bandwidth low-latency networks are being combined to be used effectively for scientific and technical computing. On suitable classes of algorithms, outstanding performance can be achieved with power efficiency significantly higher than traditional architectures, considerably minimising overall times-to-solution of high-end scientific problems. An efficient usage of such hybrid architectures is not a trivial task, often requiring a full redesign and refactoring of the code in order to expose those features that can be optimally exploited by a given hardware component, as massively parallel algorithmic components, that can be effectively accelerated by the GPU, or overlapping computation and communication in order to scale on thousands of CPUs. Traditional parallel programming approaches, based on the message-passing paradigm, are no longer sufficient to fulfil this task, and innovative solutions need to be addressed.

During the two weeks of the school a number of such solutions will be presented. The MPI, OpenMP, CUDA, OpenACC programming models will be introduced, together with scientific libraries and efficient I/O and data visualisation solutions and optimisation techniques. Their effective combined usage, in order to achieve an ideal exploitation of large hybrid architectures, will be discussed and experimented with extensive practical and exercise lab sessions to help clarify and consolidate the theoretical material presented.

The following topics will be covered:

  • Message passing programming model and the MPI standard
  • Shared memory systems and the OpenMP programming model
  • Scientific libraries
  • MPI I/O and other solutions for efficient data reading and writing
  • GPU programming with CUDA and OpenACC
  • Code optimization
  • Scientific visualisation

###APPLICATION PROCESS If you meet all requirements, please download the application form (http://goo.gl/hO38AD) fill it in, and send it to [email protected] no later than Sunday, April 11, 2016. Thirty applicants will be selected to participate in the Summer School. You will be notified about acceptance or rejection of your application by Monday, April 27, 2016. Those candidates whose application successfully passes the selection process will receive further information regarding the registration.

###PARTICIPATION FEE

  • Rate for academic staff, PhDs and university students: CHF 1’300
  • Rate for non-academic professionals: CHF 2’600
  • Fees include:
  • Eight full-day and one half-day intensive courses based on theoretical and practical coaching in a group of 30 participants
  • Learning materials introduced by instructors
  • A visit to the Swiss National Supercomputing Centre (CSCS)
  • Double room accommodation together with another participant throughout the duration of the course (eleven nights from Sunday, July 17 through Thursday, July 28, 2016)
  • Breakfast, lunch and dinner during weekdays
  • Breakfast during the weekend

Participants can book additional nights at their expense before or after the event at the same favorable rate of CHF 118 for a room with double occupancy, or CHF 99 for a room with single occupancy.


Graduate Student Modeling Workshop (IMSM 2016)

The 22nd Industrial Mathematical & Statistical Modeling (IMSM) Workshop for Graduate Students will take place at North Carolina State University, 17-27 July 2016. The workshop is sponsored by the Statistical and Applied Mathematical Science Institute (SAMSI) together with the Center for Research in Scientific Computation (CRSC) and the Department of Mathematics at North Carolina State University.

The IMSM workshop exposes graduate students in mathematics, engineering, and statistics to exciting real-world problems from industry and government. The workshop provides students with experience in a research team environment and exposure to possible career opportunities. On the first day, a Software Carpentry bootcamp will bring students up-to-date on their programming skills in Python/Matlab and R, and introduce them to version control systems and software repositories.

Local expenses and travel expenses will be covered for students at US institutions. The application deadline is 15 May 2016. Information is available at http://www.samsi.info/IMSM16 and questions can be directed to [email protected]

With best regards, Ilse Ipsen, Pierre Gremaud, and Ralph Smith


Submissions for next SIAM-LA digest

The next SIAM-LA Digest is due to be sent out on May 02, 2016. Please send any postings for the next Digest to siam-la at siam.org. Only SIAG/LA members may submit postings. To contact the list owner, send an email to siam-la-owner at siam.org.