NA Digest Sunday, June 10, 2001 Volume 01 : Issue 23

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

Submissions for NA Digest:

Mail to

Information about NA-NET:

Mail to

URL for the World Wide Web:

From: David Bailey <>
Date: Mon, 04 Jun 2001 12:22:03 -0700
Subject: Quad-double and Double-double Precision Software

I am pleased to announce new extended precision arithmetic software. It
provides the following data types from either C or Fortran programs:

double-double 106 mantissa bits, or approx. 32 decimal digits
quad-double 212 mantissa bits, or approx. 64 decimal digits

In addition to the basic operations, numerous transcendental functions,
including sqrt, sin, cos and exp, are also included. This software is
slower than ordinary double precision, but is substantially faster than
using arbitrary precision packages for these precision levels.

This subroutine library is coupled with C++ and Fortran-90 translation
modules, which use operator and function name overloading, to
automatically perform the task of translating C or Fortran code to use
the library. Thus in most cases, only minor modifications need to made
to existing C or Fortran program to utilize these precision types. Here
is an example of a typical Fortran-90 program that uses the package:

program testdd
use qdmodule
type (qd_real) a, b, c
a = -1.d0
b = 2.d0
c = acos (a) * sqrt (b)
call qdwrite (6, c)

This program produces the result
which is pi*sqrt(2) correct to 63 decimal digits.

This software was written by Yozo Hida of U.C. Berkeley, based on
earlier work by myself and Sherry Li of Lawrence Berkeley National
Laboratory. The software and documentation is available at the web site:

We would greatly appreciate any comments or bug reports by users of any
of this software. The above site also includes software for
multiprecision (arbitrary precision) arithmetic. The multiprecision
software will be revised for increased performance and usability in the
next few months.


David H Bailey "Computo ergo sum."
Lawrence Berkeley National Lab Tel. 510-495-2773
Mail Stop 50B-2239 Email:
Berkeley, CA 94720


From: Robert van de Geijn <>
Date: Sat, 9 Jun 2001 12:10:45 -0500
Subject: New Linear Algebra Development Environment

The Formal Linear Algebra Methods Environment (FLAME) is a new
methodology and library that supports the derivation and
implementation of (dense) linear algebra kernels like those provided
by the BLAS and LAPACK. This environment provides

- a recipe for derivation of families of algorithms. For each
algorithm in the family, the recipe leads to a statement of the
algorithms that hides indexing details. Moreover, the recipe provides
a constructive proof of the correctness of the algorithm.

- a library of C routines for the implementation of the algorithms.
Since the correctness of the algorithm does not necessarily lead to a
correct implementation due to indexing errors, we provide a library of
routines that allows the algorithm to be translated to code in a way
that minimizes the opportunity for the introduction of bugs.

- a collection of LaTeX commands and environments for writing
papers about the implementations.

- FLAME implementations are easily converted to parallel
implementations using PLAPACK.

- yes, we do have a FORTRAN interface as well.

- yes, our performance is competitive with the leading brand.

The methodology has been validated through an implementation of a
basic linear algebra library, including the BLAS, and LU, QR, and
Cholesky factorization based solvers. Moreover, the environment was
used as part of a class here at UT-Austin, where undergraduates with
limited background in linear algebra implemented a range of linear
algebra operations by following the recipe. Their projects have been
assembled into a coherent document:

John A. Gunnels and Robert A. van de Geijn, Developing Linear Algebra
Algorithms: A Collection of Class Projects, FLAME Working Note #3, The
University of Teas at Austin, Department of Computer Sciences.
Technical Report TR-2001-19. May 2001.

We welcome comments from the linear algebra community.

For more information, visit

Robert van de Geijn


From: Yair Censor <>
Date: Sun, 10 Jun 2001 06:46:38 +0200 (IST)
Subject: New Book on Parallel Algorithms in Optimization

We are pleased to announce the appearance of this new book (and apologize
if you happen to be exposed to multiple copies of this message):


D. Butnariu, Y. Censor, and S. Reich (Editors),
Elsevier Science Publishers, Amsterdam, The Netherlands, 2001.

For further details, including table of contents and bibliographic and
ordering information, please go to:

An excerpt from the book description on Elsevier's Internet site:

"..... Thus, the editors of this volume, with the support of the Israeli
Academy for Sciences and Humanities, took the initiative of organizing a
Workshop intended to bring together the leading scientists in the
field. The current volume is the Proceedings of the Workshop representing
the discussions, debates and communications that took place. Having all
that information collected in a single book will provide mathematicians
and engineers interested in the theoretical and practical aspects of the
inherently parallel algorithms for feasibility and optimization with a
tool for determining when, where and which algorithms in this class are
fit for solving specific problems.........".


From: Anshul Gupta <>
Date: Tue, 5 Jun 2001 14:18:03 -0400 (EDT)
Subject: Watson Sparse Matrix Package

64-bit versions of WSMP (Watson Sparse Matrix Package) are now available
and support both 4-byte (int) and 8-byte (long) integers. WSMP is a
high-performance direct solver for sparse linear systems. Currently,
it can solve symmetric sparse systems in serial, multithreaded, and
message-passing modes without pivoting and unsymmetric sparse systems in
serial and multithreaded parallel modes with pivoting on RS6000/SP parallel
computers. The new 64-bit version would allow users to solve very large
sparse systems on a single large SMP RS6000 system with more than 2 GB
of memory without using multiple MPI processes.

Please visit the WSMP web site for more details, documentation, and
technical papers describing WSMP's solvers and performance results.

Anshul Gupta
IBM T. J. Watson Research Center


From: Paul Oppenheimer <peo@ERC.MsState.Edu>
Date: Tue, 5 Jun 2001 13:14:27 -0500 (CDT)
Subject: Numerical Quadrature for Given points in Semi-infinite Domain

We would appreciate any information (papers, web links, subroutines,
etc.) concerning numerical quadrature schemes for a
semi-infinite domain [0,\infty) or (-\infty,0] under the following

1. The analytic form of the function to be integrated is not
2. The function values are known only for a given set of
points (which in general are not special points such
as those for Guassian Quadrature) in the domain.

Paul Oppenheimer
Engineering Research Center Mississippi State University


From: Jack Dongarra <>
Date: Thu, 07 Jun 2001 15:05:36 -0400
Subject: Survey of Freely Available Linear Algebra Software

I'm updating a survey I have been doing for the past few years on freely
available software for the numerical solution of linear algebra problems.&nbsp;
Of interest is software for high-performance computers that's available,
in source form, on the web for solving problems in numerical linear algebra,
specifically dense, sparse direct and iterative systems and sparse iterative
eigenvalue problems. The current list is at:
Please let me know about updates and corrections.
I'll post the updated list in a few weeks.
Thanks, Jack


From: Al Thaler <>
Date: Fri, 8 Jun 2001 12:20:22 -0500
Subject: Computational & Algorithmic Representations Of Geometric Objects

NSF and DARPA announce a joint program solicitation called

The Division of Computer and Communications Research (C-CR)and the
Division of Mathematical Sciences (DMS) of the National Science
Foundation (NSF) and the Defense Sciences Office of the Defense
Advanced Projects Research Agency (DARPA) plan to support research
and development teams focusing on mathematical and computational
innovations relevant to representation and computational manipulation
of geometrical objects. These awards will be administered by NSF.
Areas of specific interest include:

Computational Topology and Geometry
Computational and Geometric Cartography, including Spatial Statistics
Geometric aspects of Graphics and Computer-Aided Design (CAD)

We particularly encourage proposals for efforts involving
collaborations of experts in the mathematical and computational
sciences with other scientists, engineers, and practitioners
representing diverse application areas.

Proposals for incremental improvements or ongoing efforts will not be
selected for funding. This activity is particularly intended to
encourage new collaborations.
Two types of projects are anticipated:

1. Incubation grants: Standard awards of up to $100K for one or
two years to develop collaborations of mathematical scientists,
computer scientists, and application experts relevant to the
objectives of the CARGO program.

2. Team grants: Continuing awards of up to three years at $200-
$300K per year to support teams of investigators.

Please see the Solicitation for details: it is available at

Cognizant Program Officers:

- Wm Randolph Franklin, Numeric, Symbolic, and Geometric
Computation, Program Director, CISE/C-CR, 1145S, telephone: (703)
292-8912, e-mail:

- Alvin I Thaler, Mathematical Sciences, Program Director,
DMS/MPS, 1025N, telephone: (703) 292-4863, e-mail:

- Douglas Cochran, Applied & Computational Mathematics,
Program Manager, DARPA/DSO, telephone: (703) 696-2287, e-mail:


From: Charles Romine <>
Date: Wed, 06 Jun 2001 09:46:16 -0400
Subject: Change of Address for Chuck Romine

I've just joined the MICS office at DOE. I can now be reached at:

Charles Romine, SC-31
U.S. Department of Energy
19901 Germantown Road
Germantown, MD 20874-1290
Tel: (301) 903-5152
Fax: (301) 903-7774


From: Peter Deuflhard <>
Date: Thu, 7 Jun 2001 12:27:23 +0200
Subject: EMS-SIAM Conference in Berlin

Financial support for young EU scientists
1st EMS-SIAM Conference
"Applied Mathematics in our Changing World"
2 - 6 September 2001, Berlin

Complete information on the conference is available on the website

Requirements for

- young researchers (from EU and Associated States)

Letter of application stating the reasons for attending
the conference, i.e. with reference to study
and research fields

Short curriculum vitae

Letter/s of recommendation from Dean or department head

List of publications

Type and title of conference contribution, with an abstract
not exceeding 75 words

Applications will be reviewed by the Local Organizing
Committee. Funds may then be available from the
German Scientific Foundation, from the European Commission,
from the European Mathematical Society, EMS, and possibly

However, full financial support cannot be guaranteed.

You can send the above requirements to the Conference Office per FAX:

+49 (30) 841 85-107

Deadline for submissions:

June 30, 2001 for EU-Member states or EU associated states


From: Jochen Hittler <>
Date: Fri, 08 Jun 2001 08:47:38 +0200
Subject: Workshop in Hohenwart on Chemical Engineering and Biotechnology

Workshop on
Modelling and Computation in
Chemical Engineering and Biotechnology

October 1-4 2001
Hohenwart Forum

GAMM Fachauschuu Scientific Computing, IWR and STZ Technical Simulation

Research Network WiR Baden-Wurttemberg, Universitat Heidelberg IWR and IfI


Modelling and simulation in chemical engineering and biotechnology is
developing very rapidly and is gaining more and more importance for
under-standing the relevant phenoma and the design of processes in
science and industry. This workshop provides a forum to discuss recent
advances and innovations in the field. Leading international
experts will present new developments and concepts. The workshop will
bring together people from Chemical and Process Engineering, Chemistry,
Biology, Applied Mathematics and Computer Science and foster the
collaboration between university and industry.

Conference topics
computational fluid dynamics
simulation of reactive flows
reaction diffusion problems
computer aided process design and control
porous media flow
molecular design
protein folding
population dynamics

P. Bastian, Heidelberg
E. Gilles, Magdeburg
W. Jager, Heidelberg
J. Lengeler, Osnabruck*
E. Wintermantel, Zurich *
G. Wittum, Heidelberg
* tentative

Invited Speakers

P. Deuflhard, Berlin
W. Marquardt, Aachen*
L. Petzold, Santa Barbara
D. Ramkrishna, Purdue
S. Schuster, Berlin
C. Schutte, Berlin
J. Smith, Heidelberg
R. Takors, Julich
W. Wiechert, Siegen
* tentative


From: Chen Zhi-ming <>
Date: Wed, 6 Jun 2001 11:53:33 +0800 (CST)
Subject: Faculty Position at Chinese Academy of Sciences

"Hundred Talent Program" Position
Institute of Computational Mathematics
Academy of Mathematics and System Sciences
Chinese Academy of Sciences

Applications are invited for one full professor position in the field
of numerical analysis and/or scientific computing. This three-year
position is part of the special "Hundred Talent Program" of the Chinese
Academy of Sciences (CAS) aimed to attract outstanding Chinese
Scholars to work in CAS. A research grant of two million Chinese Yuan
plus housing benefits will be awarded. It is expected that the successful
applicant will lead a group in the Institute of Computational
Mathematics to work on important application-oriented problems which
scientifc computing plays a decisive role. Applicants are required to
have at least two-year research experience after having earned a
Ph.D. degree. The demonstration of excellence in researh is the most
important for a successful application. The deadline of application is
July 15, 2001.

Please contact

Prof. Yu Dehao
Vice-Director of the
Institute of Computational Mathmatics
Chinese Academy of Sciences
Tel: 86-10-62553888, Fax: 86-10-62542285

for further infomation. Related information can also found at:


From: Levent Gurel <>
Date: Sat, 9 Jun 2001 19:01:00 +0300
Subject: Postdoctoral Position at Bilkent University, Ankara, Turkey

Postdoctoral Position at Bilkent University, Ankara, Turkey

Research involves fast solution algorithms for computational
electromagnetics (CEM). In particular, fast intermediate-frequency
(resonance-region) solvers, such as the fast multipole method (FMM), and
high-frequency solvers employing asymptotic techniques will be considered.

Applications are sought immediately and will be accepted until the position
is filled, with no specific deadline.
Modern setting, friendly environment, satisfactory compensation package. For
more info, please see For applications and
further details, please contact Prof. Levent Gurel, Dept. of Electrical and
Electronics Engineering, Bilkent University, TR-06533, Ankara, Turkey,
Phone: +90-312-290 2096, Fax: +90-312-290 2439 (Private), e-mail:


From: Claire Poulin <poulin@CERCA.UMontreal.CA>
Date: Mon, 04 Jun 2001 11:51:13 -0400
Subject: Postdoctoral Position at CERCA


CERCA seeks applicants for a postdoctoral fellowship starting in September
2001 in the area of error estimation and mesh adaptation for the numerical
solution of compressible flows. The ideal incumbent has strong experience
in the development and programming of finite-volume methods on unstructured
meshes, especially Jameson and Roe type schemes applied to the solution of
Euler and Navier-Stokes systems of equations.

A working knowledge of error estimation and/or mesh adaptation for these
equations is an important asset. C++ programming skills are
desirable.Proficiency of French, although not compulsory, will be a
selection criteria. The fellowship is offered for one year. Renewal for a
second year is conditional upon the availability of funding.

CERCA is a not-for-profit corporation established in 1991 with a mission to
support university research of interest to industry. CERCA brings together
professors, research professionals, postdoctoral fellows and students with
a common interest in the development of numerical methods aiming to resolve
a variety of science and engineering problems.

CERCA fulfils its mission by forging partner-client relationships between
university, government and industry research scientists and by providing
them with the means of helping one another to train highly qualified
personnel, to develop emerging technologies, to resolve technical problems
and to share the knowledge resulting from R&D performed worldwide.

Those interested should send their CV to :

Dorothee Bonnet
5160, Decarie Blvd., Suite 400
Montreal, Quebec H3X 2H9

Applications will be accepted until July 31, 2001 and after until the=20
position is filled.


From: R.M.M. Mattheij <>
Date: Tue, 5 Jun 2001 08:19:32 +0200
Subject: Postdoctoral Position at Technische Universiteit Eindhoven

TMR project "Differential Equations in Industry and Commerce" (DEIC)
At the Scientific Computing Group a position is available for a postdoc, who
will work on problems coming from industry. There is a large variety of
interesting problems where can be worked on, ranging from materials, like
glass and polymers to plasmas. The common denominator is modelling,
mathematical analysis and computational methods. In particular those who are
keen on working on modelling are particular invited to apply.
The group consists of some 25 people (tenured staff, postdocs and PhD
students) and is co-operating with many engineering groups within the
university as well as industrial companies.
The present position is part of the European TMR project "Differential
Equations in Industry and Commerce" (DEIC). In this 6 centres are
co-operating on subjects from industrial mathematics. By EU regulation, only
inhabitants of the EU or EU associates may apply.
For further information you are invited to consult the web:
If you are interested, please write a CV with your interests to
Prof. Dr. R.M.M. Mattheij
Dept of Mathematics
Technische Universiteit Eindhoven
P.O. Box 513
5600 MB Eindhoven
The Netherlands
or by e-mail to the address:


From: Alexandre Ern <>
Date: Tue, 05 Jun 2001 16:15:19 +0200
Subject: Research Position at CEMAGREF

Poste d'ingenieur de recherche au CEMAGREF

Le CEMAGREF recrute un ingenieur de recherche dans la specialite Analyse
numerique - Calcul scientifique. L'affectation se fera dans l'unite de
recherche "Ouvrages pour le drainage et l'etancheite" basee a Antony(92).

L'ingenieur de recherche devra formaliser en termes mathematiques les
problemes poses par les specialistes des processus (essentiellement
physiques) de l'unite de recherche et proposer des strategies numeriques
pour les resoudre. Les problemes physiques a traiter, transversaux aux
equipes de l'unite, concernent principalement les transferts d'eau et de
solutes dans les milieux poreux heterogenes complexes. Responsable de
l'activite "analyse numerique" de l'unite, il devra encadrer des
stagiaires participant a l'implementation informatique des modeles et,
en collaboration avec les scientifiques de l'unite, participer a
l'encadrement des travaux de these en modelisation. Il aura egalement
pour mission de developper un reseau de partenaires universitaires.

Le profil de poste est disponible a l'adresse suivante :

Tous les renseignements concernant le dossier d'inscription au concours
et l'organisation de ce dernier se trouvent a :

Dates limites : retrait du dossier jusqu'au 14 Juin, dep=F4t des
candidatures jusqu'au 21 Juin.


From: Iain Duff <>
Date: Thu, 7 Jun 2001 09:19:24 +0100 (BST)
Subject: Contents, IMA Journal of Numerical Analysis

Contents of Volume 21, Number 3.

K. Miller and M. J. Baines
Least squares moving finite elements
pp 621-642

Z. Zhang, N. Yan and T. Sun
Superconvergent derivative recovery for the intermediate finite
element family of the second type.
pp 643-665

V. Styles
Error estimates for a finite-difference approximation of a mean
field model of superconducting vortices in one-dimension.
pp 667-701

R. Zvan, P. A. Forsyth, and K. R. Vetzal
A finite volume approach for contingent claims valuation.
pp 703-731

B. M. Garay and P. L. Simon
Numerical flow-box theorems under structural assumptions.
pp 733-749

L. Gr\"une
Persistence of attractors for one-step discretization of ordinary
differential equations.
pp 751-767

D. Potts and G. Steidl
A new linogram algorithm for computerized tomography.
pp 769-782


End of NA Digest