NA Digest Sunday, September 3, 2006 Volume 06 : Issue 36

Today's Editor:
Tamara G. Kolda
Sandia National Labs

Submissions for NA Digest:

Mail to

Information via email about NA-NET:

Mail to


From: "Stephen Vavasis" <>
Date: Thu, 31 Aug 2006 00:21:24 -0400
Subject: Change of address for Stephen Vavasis

Dear Colleagues,

After 17 years at Cornell, I am moving to University of Waterloo effective
Sep 1, 2006. My new address is

Stephen Vavasis
Department of Combinatorics and Optimization
MC 6054
University of Waterloo
200 University Avenue West
Waterloo, Ontario N2L 3G1
+1-519-888-4567 x32130
vavasis at math dot uwaterloo dot ca

My new web-page is not posted yet. You can find it by googling me in a
couple of weeks.

Yours truly,
Steve Vavasis


From: "Robert A. van de Geijn" <>
Date: Fri, 01 Sep 2006 17:35:21 -0500
Subject: FLAME libraries available under LGPL

Title: FLAME libraries available under LGPL

We are happy to announce the beta release of the FLAME libraries, libFLAME.

What is FLAME?

FLAME is a methodology for developing dense linear algebra libraries
that is radically different from the LINPACK/LAPACK approach that has
been pursued since the 1970s. libFLAME denotes the FLAME libraries that
have resulted from the effort.

libFLAME supports

- Instruction. The FLAME API is ideally suited for instruction of
high-performance linear algebra libraries at the undergraduate and
graduate level.

- Library development. The FLAME API is as much an infrastructure for
developing libraries as it is a set of libraries in their own right.

- New application development. Those who are embarking on new
application development that require a library for dense linear algebra
operations are advised to code directly using the FLAME API.

- Applications that utilize LAPACK. Those who currently use the LAPACK
library can often benefit from better performance by linking to the
libFLAME via the LAPACK-to-FLAME interface.

For addition information, visit the FLAME home page

For the FLAME Team,
Field Van Zee
Robert van de Geijn


From: "Hesthaven, Jan Sickmann" <>
Date: Tue, 29 Aug 2006 10:07:54 -0400
Subject: SPDE Workshop at Brown - Oct 20-22, 2006

Final Announcement of a DARPA and AFOSR Sponsored Workshop entitled

Advances and Challenges in the Solution of Stochastic
Partial Differential Equation

October 20-22, 2006

Brown University, Providence, RI

As methods for the deterministic solution of partial differential equations
continue to mature, the impact of uncertainty, lacking data, and noise emerge
as a fundamental challenge in the need to accurately and reliably model
complex systems.

It is the purpose of the workshop to focus on the above mentioned issues
in more detail and, by bringing together experts in various fields, initiate a
discussion on how to effectively address these significant challenges. The
format of the workshop will be as a number of 45 min long invited lectures
with discussions following that. The lecture are intended to delineate the
problems encountered as well as presenting a variety of existing and emerging
techniques for dealing with stochastic phenomena, in particular with relation
to stochastic partial differential equations. It will be a priority to attempt
to understand the advantages and limitation of the different techniques to
suggest directions in which research should be directed.

Topics that we hope will be discussed in detail during the workshop include
particle methods, Wiener chaos techniques, MC and MCMC methods,
numerical methods for stochastic ordinary and partial differential equations etc
There will also be several speakers discussing applied aspects and large
contemporary problems with a genuine need for efficient and accurate ways of
dealing with stochastic effects and uncertainty.

A prelimenary list of speakers include

B. Birnir, UCSB
A. Chorin, UC Berkeley
D. Crisan, Imperial(UK)
R. Ghanem, USC
J. Glimm, SUNY-SB
T. Hou, CalTech
G. Karniadakis, Brown
P. Kotelenez, Case Western
H. Kushner, Brown
S.V. Lototsky, USC
J. Mattingly, Duke
B. Oksendal, Oslo(N)
G. Papanicolau, Stanford
B. Rozovsky, USC
A. Stuart, Warwick(UK)
D. Talay, INRIA(F)
R .Temam, Indiana
J. Tribbia, NCAR
E. Vander-Eijden, NYU

Details about hotels, transportation, prelimenary registration etc can be
found at

If you have specific questions regarding this workshop, please feel free to
contact the workshop secretary

Ms Janice D'Amico
(401) 863-2358

Best regards

The organizing comittee

Paul Dupuis (Brown), David Gottlieb (Brown)
Jan S. Hesthaven (Brown), Boris Rozovsky (Brown)


From: "Sheng, Qin" <>
Date: Thu, 31 Aug 2006 15:38:52 -0500
Subject: Second Announcement of Special Issue on Splitting Methods

Second Announcement of Special Issue on Splitting Methods

The International Journal of Computer Mathematics will publish a special issue
(Baylor University) and A.Q.M. Khaliq (Middle Tennessee State University).
Different splitting methods have played important roles in the numerical
solution of differential equations. Since the pioneer work on ADI and
exponential splitting formulae, many new concepts and strategies have been
developed and utilized. The aim of this special issue is to highlight the
new developments in the area. It will contain articles presenting the latest
trends and research results in topics including, but not limited to:

* splitting for higher efficiency and accuracy
* splitting for non-linear differential equations
* stability and convergence of splitting methods
* iterative and adaptive splitting methods
* splitting methods in parallel and quantum computations

Research papers are solicited for this special issue. Each submitted paper
should be between 10 and 16 pages and will be refereed according to IJCM
policies ( ).
Submit a PDF or PS version of the complete paper to either of the guest

Professor Qin Sheng
Department of Mathematics
Baylor University
Texas, USA


Professor Abdul Q. M. Khaliq
Department of Mathematical Sciences
Middle Tennessee State University
Tennessee, USA

Deadline for submission of full papers: October 31, 2006.
Notification of acceptance: February 28, 2007.
Expected publication: Summer, 2007.


From: "John G. Lewis" <>
Date: Thu, 31 Aug 2006 11:08:14 -0700
Subject: Addendum to Announcement of Positions at Cray Inc.

The instructions on how to submit applications was truncated from last
week's announcement. Interested parties should visit
The "Opportunities Search" link will open a job search mechanism in a
new window. Enter requisition number 2268 for the group leader
position or 1300 for the FFT software engineering position to see a
complete job description and a link to submit an online job


From: "Charlotte Smith" <>
Date: Thu, 31 Aug 2006 16:34:13 +0100
Subject: Lecturerships in Computational Biology at the University of Oxford


Mathematical, Physical and Life Sciences Division
in association with Linacre College and Green College


Applications are invited for two University Lecturerships in the field of
Computational Biology. These are permanent posts associated with either
Linacre College or Green College, under arrangements described in the further
particulars. The combined University and College salary will be on a scale up
to £50,589 per annum.

The successful candidate will join the Computational Biology Group within the
Computing Laboratory, which is led by Professor David Gavaghan. Applicants
should have an internationally leading research track record in any area
complementary to the current interests of the group. These focus on the
mathematical modelling, numerical analysis, software engineering and
(parallel) scientific programming research issues arising within computational
biology research. In this particular instance, applications are particularly
encouraged from those with research interests relevant to the theoretical
underpinning of the group's work in mathematical modelling, numerical methods
and scientific programming.

The lecturers will be affiliated with the Computing Laboratory. The appointees
are expected to have an excellent research record in Computational Biology,
and are required to do first class research in this and related areas at
Oxford. The applicants are expected to further strengthen and build the
existing research group and to cooperate towards this goal with Professor
Gavaghan. Moreover, the applicants should be willing and able to attract and
carry out funded research projects.

Full details of the posts, the application procedure and information about the
University, Linacre College and Green College are given in the Further
Particulars which are available from or via
email request to

The closing date for both applications is 29th September 2006.


From: Frederic Nataf <>
Date: Thu, 31 Aug 2006 20:49:23 +0200
Subject: Postdoc Position at French Petroleum Institute / Paris VI University

Postdoc Position at French Petroleum Institute / Paris VI University
Parallel Linear Solvers for Large Sparse Systems arising from
Petroleum Industry

The French Petroleum Institute is seeking a scientist or engineer
with a strong background in computer science and high performance
computing for the solution of large sparse linear systems.
The applicant will work on the development of a parallel C++ platform
for algebraic domain decomposition methods. Based on sparse
distributed matrices data structures the applicant will develop a new
computational framework for interface conditions and coarse grid
correction. Different strategies will be studied and tested on
industrial problems for reservoir and basin modeling simulations.
The candidate must have a PhD degree from a recognized university in
engineering, computer sciences or applied mathematics. Previous
experience on parallel linear algebra and strong computational skills
in C++ will be an asset.

Duration : 12-18 months
Location : French Petroleum Institute and University Paris VI, Paris
Salary (gross wage) : 2800 Euros/month (3200 Euros/month for foreign
students from foreign universities)

Send your application (CV + references) to

Frederic Nataf
Jacques Louis Lions Laboratory, University Paris VI


From: Espen Robstad Jakobsen <>
Date: Tue, 29 Aug 2006 15:22:01 +0200 (MEST)
Subject: PhD position in Numerical Analysis and Applications in Finance at NTNU


A PhD grant is available at the Department of Mathematical Sciences
at the Norwegian University of Science and Technology (NTNU). The grant,
which is for three years, is funded by the Research Council of Norway. The
grant period may be extended to four years in exchange for teaching duties
at the department.

The PhD project focuses on development and analysis (theory) of
numerical methods for integro-partial differential equations and
applications in finance. Related theoretical topics as boundary values and
regularity of solutions may be considered additionally. The PhD
candidate will participate in the activities of the Integro-PDEs:
Numerical methods, Analysis, and Applications project which is based at
NTNU and at the University of Oslo.

More information can be found on the web page

For more information, please contact:

Associate Professor Espen R. Jakobsen
phone: +47 73 59 35 12

Closing date: 2006-09-27.


From: Charles Augarde <>
Date: Thu, 31 Aug 2006 10:02:21 +0100
Subject: Research Associate, Computational Geomechanics, Durham Univ.

A Research Associate is required to work on an exciting project to
develop a novel numerical method for application in geotechnics and many
other areas of computational mechanics. The post will suit someone with
(or about to get) a PhD in civil or mechanical engineering or applied maths.

The School of Engineering at Durham has a thriving group working on
computational techniques applied to civil and mechanical engineering,
particularly geotechnics. This post will assist on an exciting
EPSRC-funded project to develop and extend a fledgling numerical method.
The method to be developed is based on a combination of a meshless
method and a scaled boundary method and removes many of the drawbacks of
traditional finite element methods. The post will involve background
investigation of meshless and scaled boundary methods, coding in fortran
90/95 working from the existing Matlab implementation, testing and
finally validation on a challenging geotechnical problem.

The Associate will work closely with Dr Charles Augarde and Prof. Roger
Crouch at Durham, and with a Project Student to be appointed at the same
time. The Associate will also benefit from the input of two
international experts in computational mechanics who will be
collaborating. There is also the possibility of the Associate spending a
period working at UWA in Perth, Australia as part of the project. The
post will suit someone with a keen interest in programming, numerical
methods and geotechnics.

Full details at vacancy ref 1556
Also on
Closing date 21/09/2006.


From: Chi-Wang Shu <>
Date: Fri, 01 Sep 2006 17:00:38 -0400
Subject: Contents, Journal of Scientific Computing

Journal of Scientific Computing

Volume 28, Numbers 2-3, September 2006

Special issue in honor of Professor David Gottlieb on
the occasion of his 60th birthday

Saul Abarbanel, pp.121-123.

Non-Linear PML Equations for Time Dependent Electromagnetics
in Three Dimensions
S. Abarbanel, D. Gottlieb and J. S. Hesthaven, pp.125-137.

A Mathematical Approach in the Design of Arterial Bypass
Using Unsteady Stokes Equations
Valery Agoshkov, Alfio Quarteroni and Gianluigi Rozza,

One-sided Post-processing for the Discontinuous Galerkin
Method Using ENO Type Stencil Choosingand the Local Edge
Detection Method
Rick Archibald, Anne Gelb, Sigal Gottlieb and Jennifer Ryan,

Homogenization of Periodically Varying Coefficients in
Electromagnetic Materials
H.T. Banks, V.A. Bokil, D. Cioranescu, N.L. Gibson, G. Griso
and B. Miara, pp.191-221.

Enhanced Coercivity for Pure Advection and Advection–Diffusion
Claudio Canuto, pp.223-244.

Prediction from Partial Data, Renormalization, and Averaging
Alexandre J. Chorin, Ole H. Hald and Raz Kupferman, pp.245-261.

A Boundary Integral Equation Method for Photonic Crystal Fibers
Min Hyung Cho, Wei Cai and Tsing-Hua Her, pp.263-278.

Adaptive Edge Detectors for Piecewise Smooth Data Based on
the minmod Limiter
A. Gelb and E. Tadmor, pp.279-306.

Recovering High-Order Accuracy in WENO Computations of
Steady-State Hyperbolic System
Sigal Gottlieb, David Gottlieb and Chi-Wang Shu, pp.307-318.

The Shifted Box Scheme for Scalar Transport Problems
Bertil Gustafsson and Yaser Khalighi, pp.319-335.

Pade-Legendre Interpolants for Gibbs Reconstruction
J.S. Hesthaven, S.M. Kaber and L. Lurati, pp.337-359.

On Parabolic Boundary Layers for Convection–Diffusion
Equations in a Channel: Analysis and Numerical Applications
Chang-Yeol Jung and Roger Temam, pp.361-410.

Denoising by BV-duality
Stefan Kindermann, Stanley Osher and Jinjun Xu, pp.411-444.

Gibbs Phenomena
Peter D. Lax, pp.445-449.

L-infinity-Stable Approximation of a Solution to Div(Y) = f
for f in L2 in Two Dimensions
Yvon Maday, pp.451-458.

Turbulence Effects on Kinetic Equations
Steven A. Orszag, Hudong Chen, Sauro Succi, Jonas Latt and
Bastien Chopard, pp.459-466.

On the Method of Difference Potentials
V. Ryabenkii, pp.467-478.

Digital Flight: The Last CFD Aeronautical Grand Challenge
M.D. Salas, pp.479-505.

Simulating Plasma Microwave Diagnostics
J.S. Shang, pp.507-532.

Numerical Study of a Modified Time-Stepping theta-Scheme for
Incompressible Flow Simulations
Stefan Turek, Ludmila Rivkind, Jaroslav Hron and Roland
Glowinski, pp.533-547.

Numerical Methods and Nature
E. Turkel, pp.549-570.


From: Claude Brezinski <>
Date: Fri, 01 Sep 2006 09:50:56 +0200
Subject: Contents, Numerical Algorithms

Volume 42 Number 2 of Numerical Algorithms is now available
on the web site at

The following URL will take you directly to the issue:

This issue contains:

On applicability of the sparse grid method in the worst case setting
Fengbo Hang, Youming Li

PDE-constrained control using Femlab ­ Control of the Navier­Stokes equations
Thomas Slawig

On some families of multi-point iterative methods for solving nonlinear equations
Gyurhan H. Nedzhibov, Vejdi I. Hasanov, Milko G. Petkov

Efficient spectral-Galerkin algorithms for direct solution for second-order differential equations using Jacobi polynomials
E. H. Doha, A. H. Bhrawy

A numerical method for solving optimal control problems using state parametrization
H. Hashemi Mehne, A. Hashemi Borzabadi

Computing orthogonal polynomials on a triangle by degree raising
Shayne Waldron

An iterative method for the least squares symmetric solution of matrix equation
Jin-jun Hou, Zhen-yun Peng, Xiang-lin Zhang

Stable Runge­Kutta­Nystroem methods for dissipative stiff problems
I. Alonso-Mallo, B. Cano, M. J. Moreta

End of NA Digest