NA Digest Sunday, March 31, 2002 Volume 02 : Issue 13

Today's Editor:
Cleve Moler
The MathWorks, Inc.

Submissions for NA Digest:

Mail to

Information via e-mail about NA-NET: Mail to


From: Jorge More' <>
Date: Sun, 24 Mar 2002 15:09:25 -0600
Subject: Wilkinson Prize for Numerical Software


In honor of the outstanding contributions of James Hardy Wilkinson to
the field of numerical software, Argonne National Laboratory, the
National Physical Laboratory, and the Numerical Algorithms Group award
a numerical software prize of US $1000. The first prize was awarded
to Linda Petzold for DASSL at the International Conference in
Industrial and Applied Mathematics (ICIAM 91), the second prize was
awarded to Chris Bischof and Alan Carle for ADIFOR 2.0 at ICIAM 95,
and the third prize was awarded to Matteo Frigo and Steven Johnson for

The 2003 prize will be awarded at ICIAM 2003 in Sydney, July 7-11, 2003.

Rules for Submission

Each author of an entry must be at most 40 years of age on January 1, 2003.
Each entry must contain the following:

Software written in a widely available high-level programming language.

A paper describing the algorithm and the software implementation.
The paper should give an analysis of the algorithm and indicate any
special programming features.

Documentation of the software, which describes its purpose and method of use.

Examples of use of the software, including a test program and data.
A (two page) summary of the main features of the algorithm and
software implementation.

Submissions must be in English. Entries must be received by November 4, 2002.

Selection Criteria

The award will be made to the entry that best addresses all phases of the
preparation of high-quality numerical software, including

Clarity of the paper and of the software implementation and documentation;

Portability, reliability, efficiency, and usability of the software implementation;

Depth of analysis of the algorithm and the software;

Importance of application addressed by the software;

Quality of the test software.


Submissions ideally be in the form of a uuencoded, gzipped, tar archive.
Submissions should include a README file describing the contents of the
archive and Makefiles for executing the test programs.
Submissions can be sent by email to
Contact this address for further information.


From: David Bruhwiler <>
Date: Wed, 27 Mar 2002 10:30:17 -0700 (MST)
Subject: Release of QScimpl Graphics Package

Tech-X Corporation is pleased to announce the 2nd beta
release of

QScimpl --
Qt-based Scientific Modeling and Plotting Library

Based on the cross-platform C++ GUI toolkit developed
by Trolltech AS, QScimpl (pronounced "Q-simple") is a
scientific graphics package that provides additional
functionality for the rapid development of GUI's for
scientific codes that require interactive visualization
of particle or field data.

Like Qt, QScimpl is readily extensible by developers.

QScimpl supports 3-D rendering with OpenGL or Mesa,
as well as threaded operation; however, neither of
these features are required.

QScimpl is distributed with many example executables.
One of these, called "SciMovie", can animate a sequence
of images in a variety of different formats.

QScimpl comes with full source code and is free for
non-commercial use.

Please point your browser to URL
for more information.

Best regards,

David L. Bruhwiler -
Tech-X Corporation -
(303) 448-0732


From: David Bruhwiler <>
Date: Wed, 27 Mar 2002 10:35:46 -0700 (MST)
Subject: Release of OOPIC Particle in Cell Code

Tech-X Corporation is pleased to announce the beta
release of OOPIC Pro Version 3.0

OOPIC is a high-performance 2-D PIC (particle-in-cell) code,
with support for x-y (slab) and r-z (cylindrical) geometries.

OOPIC Pro is written in C++. It can be run in serial mode
with a GUI or in batch mode on many processors.

For more information, point your browser to URL

OOPIC Pro is freely available for non-commercial use on
Unix and Linux platforms, with full source code provided.

An MS Windows version of OOPIC Pro is available for beta test.

Best regards,

David L. Bruhwiler -
Tech-X Corporation -
(303) 448-0732


From: George Karypis <>
Date: Thu, 28 Mar 2002 02:11:44 -0600
Subject: ParMETIS, Graph Partitioning and Sparse Matrix Ordering Library

ParMETIS 3.0: A Parallel Graph Partitioning and Sparse Matrix Ordering

We would like to announce the release of version 3.0 of the ParMETIS
library. ParMETIS is an MPI-based parallel library that implements a
variety of algorithms for partitioning unstructured graphs, meshes, and for
computing fill-reducing orderings for sparse matrices. ParMETIS is
particularly suited for parallel numerical simulations involving large
unstructured meshes. For these computations, ParMETIS's algorithms are
based on the multilevel partitioning and fill reducing ordering algorithms
that are implemented in the widely used serial package METIS. ParMETIS
extends the functionality provided by METIS by including routines that
are especially suited for parallel computations and large scale numerical

ParMETIS provides the following four major functions:
- Partition an unstructured graph and/or mesh.
- Improve the quality of an existing partition.
- Repartition a graph that corresponds to an adaptively refined mesh.
- Compute a fill-reducing ordering for sparse direct factorization.

Obtaining ParMETIS

ParMETIS is distributed freely. Information on how to download the
source code is available on WWW at


ParMETIS has been written by George Karypis and Kirk Schloegel, at the
Computer Science Department of the University of Minnesota. If you have
any questions or problems obtaining ParMETIS, send email to


From: Joe Grcar <>
Date: Fri, 29 Mar 2002 19:06:37 -0800
Subject: Regarding Numerical Software

Regarding Numerical Software

I would like to suggest a topic for discussion: the role of this
community in developing NA software.

Wilkinson and Reinsch edited a handbook about NA software many
years ago. Their purpose was to showcase the best algorithms
(for linear algebra), and to illustrate best software practices.
An implicit purpose, I think, was to demonstrate that NA
specialists are needed.

I have the uneasy feeling that Wilkinson and Reinsch's interests
in the software aspect of NA are still important, but they are
not being pursued as well as they should be. Of course, there are
lots of ways to contribute to NA, but ultimately they have to
impact how people use floating point arithmetic. The problem is
that it is hard to see where NA research actually gets turned
into software today:

a) Only a few commercial products (imsl, maple, matlab, nag)
appear to be maintained by people trained in NA.

b) Except for a few NA bits, most scientific and engineering
software is written by scientists and engineers.

c) It is not likely that someone from another field can write a
program from what they find in NA journals, because NA papers
are very mathematical.

So where does the NA rubber meet the software road?

Perhaps we are seeing what Parlett predicted in SIAM Review
some years ago. He speculated that NA specialists would be
dispersed among the sciences; so maybe those people are doing
the programming. If so, then NA software is even more of a
concern, because most fields do not emphasize software quality
(and more to the point, algorithmic quality).

No matter what your guess about the future of NA, it does
seem that NA specialists have a responsibility to set the
standard for quality in floating point software. If not them,
then who? This will be more important as universities set up
departments of computational science, and begin to teach what
amount to curricula for NA software majors.

But just when leadership is most needed, the NA community seems
to have de-emphasized software. It would take demographic studies
to prove this, so instead I will offer some examples. Among all
the prizes (dozens) that ACM and SIAM give out, only two are for
NA software, and one isn't even awarded every year. Moreover, no
journal publishes articles about NA software practices (not SIAM,
and not ACM TOMS, which is an algorithm collection).

Even for NA specialists who want to work with software, many
hurdles make it difficult to do so.

A) There is no place to publish, and little recognition.

B) There is very little linguistically modern NA software on
which to build because it seldom gets rewritten to keep up with
evolving hardware and languages.

C) As NA research becomes more specialized, less software is
algorithmically complete. For example, today it may be clear that
method A is the best way to solve problem P. So to work on
problem P, you need to start with some software for algorithm A.
Where do you get it? The software should include all the ideas
that other people working on problem P have contributed in the
last 5 years (and it should use the latest hardware and language).

D) Some algorithms are so complicated that they cannot be
programmed by one person. (I do think there are such things.)

The hurdles faced by NA specialists who teach are even higher.
If such a person wants to write research software, their
activities are essentially limited in scope to what a single
person can do from scratch in between classes. This excludes
some kinds of problems from ever being considered, and it
precludes certain kinds of collaborations with other fields.
All of this is to the detriment of their students.

The possibility of collaborations with other fields is important
because it is at the heart of those computational science
departments we hear aobut. Scientific research is typically done
in groups. NA specialists can be members of such groups, but if
the NA specialists have no track record in software, then it is
not likely that they will be asked to participate. Some people do
maintain private software infrastructures and they make outstanding
contributions in collaboration with scientists. However, such
efforts are beyond many people's reach and are increasingly heroic:
the supply of students who program well is sporadic, and grants
rarely support continuing software projects.

So it seems to me that some effort is needed to get back to the
kind of NA that Wilkinson and Reinsch believed in, where there
was more of a balance between math and software.

I think what is most needed are some modest "institutional" changes
that over time would counteract the forces that seem to be drawing
NA specialists away from software. Here are some suggestions about
what could be done.

(1) NA journals could accept papers about software. They already
accept papers with theorems and proofs, which are the math
side, so why not accept papers about the CS side of NA

(2) There could be a continuing series like Wilkinson and Reinsch's
handbook about NA software technology. Students do not know
what modern NA software looks like, because there is none for
them to look at.

(3) There could be a classification of NA software so that journal
articles that discuss software can be indexed. There is a 70-page
classification of all math research, so why not NA software?

(4) It may be possible to develop large NA software suites in the
same cooperative, voluntary way that unix operating systems are
created, such as freeBSD, gnu, or linux.

(5) People in authority should hear that since NA is not a subset
of mathematics, so research grants awarded to NA specialists
need to have a broader scope that will enable work in [your
software project here].

(6) Although I like big machines and scientific collaborations,
I think it is an overworked paradigm to rely on them to justify
support for NA specialists, because they are not a replacement for
a basic research effort. So, my friends who are program managers
at the NSF and elsewhere: it is a disservice to craft requests-
for-proposals that entice the academic community into this type
of work without also providing alternative funding sources for
core research activities.

(7) Perhaps SIAM needs a SIG for NA software.

(8) There should be an NA software repository that is operated by
people with rotating duties, elected leaders, and public guidelines
for what is deposited. One could imagine that the software
would be free, but the documentation would be in books (see item
2) whose royalties might provide some support for the repository.

(9) Students who want to become computing professionals should be
encouraged to undertake NA software projects for masters degrees.
Supervising faculty could preregister the project for inclusion
in the software repository (see item 8) to give the student a
customer (real software projects have customers) and recognition.

(10) Standard-setting organizations need more input from the NA
community. For example, how should the c++ stl be modified to
better support NA software? This does not mean a lifetime
committee assignment. If there were more community-wide
activity in NA software, then it would be easier to make a case.

I'm not sure why I wrote this. I guess it seemed to me that a
part of NA I have always enjoyed was starting to fall by the
wayside. When I looked around for some discussion about it I
couldn't find any, which reinforced my suspicions. So these are
my thoughts. I would be interested in hearing yours.

Regards, --Joe

Joseph Grcar
Lawrence Berkeley National Laboratory
Mail Stop 50A-1148
One Cyclotron Road
Berkeley, CA 94720

(510) 495-2810 voice
(510) 486-6900 fax


From: Laurence Yang <>
Date: Sun, 24 Mar 2002 16:16:57 -0400
Subject: Workshop in Vancouver on High Performance Computing

The 4th Workshop on High Performance Scientific and Engineering
Computing with Applications (HPSECA-02)

The Renaissance Vancouver Hotel Harbourside Vancouver, British Columbia,
August 18-21 (Sunday-Wednesday), 2002


Scope and Interests:

Parallel and distributed scientific and engineering computing has become
a key technology which will play an important
part in determining, or at least shaping, future research and
development activities in many academic and industrial
branches. This special workshop is to bring together computer
scientists, applied mathematicians and researchers to
present, discuss and exchange idea, results, work in progress and
experience of research in the area of parallel and
distributed computing for problems in science and engineering

Among the main topics (but not limited to) are:

development of advanced parallel and distributed methods,
parallel and distributed computing techniques and codes,
practical experiences using various supercomputers with software such as
MPI, PVM, and High Performance Fortran, OpenMP, etc.
applications to the following areas, but not limited to:
computational fluid dynamics and mechanics
material sciences
space, weather, climate systems and global changes
computational environment and energy systems
computational ocean and earth sciences
combustion system simulation
computational chemistry
computational physics
bioinformatics and computational biology
medical applications
transportation systems simulations
combinatorial and global optimization problems
structural engineering
computational electromagnetics
computer graphics
semiconductor technology,and electronic circuits and system
dynamic systems
computational finance

Further information about the conference proceedings and registration
fee can be found by web sites:

Workshop Organizers:

Prof. Laurence T. Yang (chair)
Department of Computer Science
PO Box 5000,
St. Francis Xavier University
Antigonish, B2G 2W5, Nova Scotia, Canada

Prof. Yi Pan (Co-Chair)
Department of Computer Science,
Georgia State University
Atlanta, GA 30303 USA


From: Dynamics Days <>
Date: Mon, 25 Mar 2002 21:27:41 +0100 (MET)
Subject: Dynamics Days Europe


The deadline for abstract submission to

XXII. Dynamics Days Europe 2002
Heidelberg, Germany ---- July 15--19, 2002

is approaching on March 31. This is also a strict deadline for applications
for financial support. Details can be found on our web pages at

or in the updated ANNOUNCEMENT AND CALL FOR ABSTRACTS given in the following.


XXII. Dynamics Days Europe 2002
Heidelberg, Germany ---- July 15--19, 2002

Dynamics Days Europe is a major international conference with a long
tradition aimed at covering the entire field of dynamics and
nonlinearity. The XXIInd event in this tradition will take place in

Heidelberg, July 15-19, 2002

at the communication center of the German Cancer Research Center
on the campus of the University of Heidelberg.


The focus of the conference will be on new developments in modelling,
mathematical foundations, applications and experiments. The invited
sessions cover

Computational Physics Patterns and Waves
Dynamical Systems Nonlinear Quantum Effects
Engineering and Optimization Statistical Physics
Fluid Dynamics Stochastics and Applications
Life Sciences Structured Devices


W. Achtziger (U Erlangen) K. Mischaikow (Georgia Tech)
L. Arnold (U Bremen) M. Moeller (U Ulm)
G. Benettin (U Padova) S. Mueller (MPI Leipzig)
E. Brener (FZ Juelich) Z. Noszticzius (TU Budapest)
J. Eggers (U Essen) R. Phair (BioInformatics Services)
M. Eiswirth (FHI Berlin) H. A. Posch (U Vienna)
G. Falkovich (Weizmann) I. Procaccia (Weizmann)
T. Fukuda (Nagoya U) D. Quere (College de France)
E. D. Gilles (MPI Magdeburg) D. Ruelle (IHES Paris)
G. I. Goldburg (U Pittsburgh) K. Sacha (U Krakow)
M. Inagaki (Toyota CRDL) B. Sandstede (Ohio State)
W. Just (TU Chemnitz) B. Schmittmann (Virginia Tech)
Y. G. Kevrekidis (Princeton) H. Schomerus (MPI Dresden)
P. Kotelenez (CWRU Cleveland) C. Schuette (TU Berlin)
G. Leuchs (U Erlangen) A. K. Sood (IISC Bangalore)
M. Marek (ICT Prague) A. Stevens (MPI Leipzig)
H. Matano (U Tokyo) P. Tabeling (ENS Paris)
P. Mendes (Virginia Tech) L.-S. Young (CIMS New York)
A. Mielke (U Stuttgart) J. Zhang (CIMS New York)
A. S. Mikhailov (FHI Berlin)

ORGANIZERS: Jens Starke (U Heidelberg)
Juergen Vollmer (MPI Mainz)
SCIENTIFIC HOST: Roland Eils (German Cancer Research Center)

SPONSORS: Sonderforschungsbereich 359, Heidelberg
German Cancer Research Center, Heidelberg
Max-Planck-Institute for Polymer Research, Mainz
Deutsche Forschungsgemeinschaft
European Physical Society
Heraeus Stiftung, Hanau
Megware Computer, Chemnitz


From: Irwin Pressman <>
Date: Mon, 25 Mar 2002 17:39:12 -0500 (EST)
Subject: Discrete Math Day at Carleton

Ottawa-Carleton Discrete Math Day Saturday April 13, 2002
Carleton University (Ottawa, Canada).

This is an annual event. There will be 4 talks on April 13 by Jeff
Dinitz, Bill Pulleyblank, Lieven Vandenberghe and Herb Wilf. In addition,
there will be a Colloquium Seminar on Friday April 12, 2002 given by Luc

There is no registration fee but we would appreciate if you let us know
that you are attending the meeting by e-mailing:

Further information related to this day can be found at:


From: Heike Fricke <>
Date: Thu, 28 Mar 2002 15:10:29 +0100
Subject: Research Position at University of Tuebingen

The Sonderforschungsbereich (SFB) 382, Methods and Algorithms for
Simulating Physical Processes on Supercomputers, is looking for a

Head of a Junior Research Group (BAT Ia, German university scale)

The Junior Research Group will be established within the field of
scientific computing (e.g. computational physics, numerical mathematics,
computergraphic and visualization, high performance computing)

which will be closely associated with the SFB. Initially, the post
is available for 2 years (until June 30, 2004), renewal is possible
for altogether five years.

The appointee should have scientific experience in at least one of
the named fields. He will apply for his own project within the
SFB. Additionally to the position of the leader, two further posts
(BAT IIa) can be applied for. This project will be appraised by
the German Research Foundation. The appointment is
dependent on the success of the project's application.

Candidates will be expected to have a doctorate in a relevant
discipline. Applications and the usual documents as well as an
informative sketch of a research proposal should be directed to
the leader of the SFB 382, Prof. Dr. Hanns Ruder, Institute for
Astronomy and Astrophysics, University of T=FCbingen, Auf der
Morgenstelle 10, D-72076 T=FCbingen. The closing date for
applications is April 30, 2002.

The current SFB staff and their research interests are described

As equal opportunity employer the university welcomes applications from
suitably qualified people from all sections of the community regardless
of race, religion, gender or disability.


From: David Bruhwiler <>
Date: Wed, 27 Mar 2002 10:36:39 -0700 (MST)
Subject: Research Position at Tech-X Corporation

Applications are invited for a permanent scientific
research position at Tech-X Corporation. Tech-X
specializes in computational science, aimed at
understanding physics via modern modeling techniques
and using new software development paradigms. We
have projects in the areas of accelerator physics,
fusion plasma physics and laser-plasma interactions.
We are interested in object oriented programming,
legacy code reuse, code steering through scripting
or flexible graphical user interfaces, visualization
and parallel processing. See
for more information about the company. The successful
applicant will have a PhD in physics or a related
field, with a strong background in computing. Both
junior and senior candidates will be considered,
with preference given to those who have experience
in modeling particle beam or plasma devices. Tech-X
offers a competitive salary and benefits package.
Applicants should submit a CV and a list of three
references along with a letter of application
stating their research interests. The closing date
for applications will be June 30, 2002. The CV may
be mailed to "Physicist Job Search", Tech-X Corp.,
5541 Central Ave., Suite 135, Boulder, CO, 80301 or
emailed to or faxed to (303) 448-7756.

Best regards,

David L. Bruhwiler -
Tech-X Corporation -
(303) 448-0732


From: Thomas Hogan <>
Date: Thu, 28 Mar 2002 16:27:24 -0500 (EST)
Subject: Contents, Journal of Approximation Theory

Table of Contents: J. Approx. Theory, Volume 115, Number 1, March 2002

Ying Guang Shi
New characterizations of ratio asymptotics for orthogonal polynomials

Ushangi Goginava
On the approximation properties of Ces\`{a}ro means of negative order
of Walsh-Fourier series

Sen-Yen Shaw and Hsiang Liu
Convergence rates of regularized approximation processes

Chong Li and Renxing Ni
Derivatives of generalized distance functions and existence of generalized
nearest points

R. K. Kovacheva and J. {\L}awrynowicz
An analogue of Montel's theorem for rational functions of best

Dimiter Dryanov and Petar Petrov
Best one-sided $L^1$-approximation by blending functions of order (2,2)

M. Castro Smirnova
Convergence conditions for vector Stieltjes continued fractions

A. L. Brown, Frank Deutsch, V. Indumathi, and Petar S. Kenderov
Lower semicontinuity concepts, continuous selections, and set valued
metric projections

Ahmed Fitouhi, M. Moncef Hamza, and Fethi Bouzeffour
The $q-j_\alpha$ Bessel function

Ying Guang Shi
Bounds and inequalities for $L_m$ extremal polynomials


From: Hans Schneider <>
Date: Fri, 29 Mar 2002 17:07:01 -0600 (CST)
Subject: Contents, Linear Algebra and its Applications

Journal: Linear Algebra and its Applications
ISSN : 0024-3795
Volume : 346
Issue : 1-3
Date : 01-May-2002

Visit the journal at

An algorithmic version of the theorem by Latimer and MacDuffee for
2x2 integral matrices
A. Behn, A.B. Van der Merwe
pp 1-14

On the nonlinear matrix equation X+A^*F(X)A=Q: solutions and
perturbation theory
A.C.M. Ran, M.C.B. Reurings
pp 15-26

Global reduction to the Kronecker canonical form of a C^r-family of
time-invariant linear systems
X. Puerta, F. Puerta, J. Ferrer
pp 27-45

Convexity and the separability problem of quantum mechanical density matrices
A.O. Pittenger, M.H. Rubin
pp 47-71

More on matrix semigroup homomorphisms
D. Kokol-Bukovsek
pp 73-95

Pole-shifting for linear systems over commutative rings
M. Carriegos, J.A. Hermida-Alonso, T. Sanchez-Giralda
pp 97-107

Maximal graphs and graphs with maximal spectral radius
D.D. Olesky, A. Roy, P. van den Driessche
pp 109-130

Obtaining simultaneous solutions of linear subsystems of inequalities and duals
E. Castillo, F. Jubete, R.E. Pruneda, C. Solares
pp 131-154

On matrix differential equations and abstract FG algorithm
M. Przybylska
pp 155-175

A polynomial fit preconditioner for band Toeplitz matrices in image
P. Favati, G. Lotti, O. Menchi
pp 177-197

Coherence invariant mappings on block triangular matrix spaces
W.L. Chooi, M.H. Lim
pp 199-238

A boundary Nevanlinna-Pick problem for a class of analytic
matrix-valued functions in the unit ball
V. Bolotnikov
pp 239-260

Irreducible, pattern k-potent ray pattern matrices
J.L. Stuart, L. Beasley, B. Shader
pp 261-271

On block completion problems for Arov-normalized j"q"q-J"q-elementary factors
B. Fritzsche, B. Kirstein, M. Mosch
pp 273-291


From: Vladik Kreinovich <>
Date: Sun, 31 Mar 2002 10:39:36 -0700 (MST)
Subject: Contents, Reliable Computing

Reliable Computing
Volume 8, issue 3, 2002

Mathematical Research

Bounding Perturbations in Zeros of Nonlinear Systems
Michael A. Wolfe

Formal Solution to Systems of Interval Linear or Non-Linear Equations
Miguel A. Sainz, Ernest Gardenyes, Lambert Jorba

In Case of Interval (or More General) Uncertainty, No Algorithm Can
Choose the Simplest Representative
Gerhard Heindl, Vladik Kreinovich, Maria Rifqi

An Approach to Overcome Division by Zero in the Interval Gauss Algorithm
Jan Mayer

Intervals of Inverse M-matrices
Charles R. Johnson, Ronald L. Smith

Short communication

Rump's Example Revisited
Eugene Loh, G. William Walster

Reliable Computing
Volume 8, issue 2, 2002

Mathematical Research

Interval-Valued Finite Markov Chains
Igor O. Kozine, Lev V. Utkin

Sharp Bounds on Interval Polynomial Roots
Eldon R. Hansen, G. William Walster

Algorithms and Computations

A Parallelized Version of the Covering Algorithm for Solving
Parameter-Dependent Systems of Nonlinear Equations
Paluri S. V. Nataraj, Airani Kalathil Prakash

Interval Computation of Viswanath's Constant
Joao Batista Oliveira, Luiz Henrique de Figueiredo

Applications and Tools

Efficient Interval Linear Equality Solving in Constraint Logic
Chong-Kan Chiu, Jimmy Ho-Man Lee


Reliable Computing: Special Issue on Dependable Reasoning about
Daniel Berleant


End of NA Digest