NA Digest Sunday, January 30, 1994 Volume 94 : Issue 5

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

Submissions for NA Digest:

Mail to

Information about NA-NET:

Mail to


From: Craig Barratt <>
Date: Mon, 24 Jan 94 15:41:00 -0800
Subject: BLAS for the DEC Alpha

Does anyone know about the availablility of optimized implementations of
BLAS Level 1, 2 and 3 for the DEC Alpha processor (21064/066/068)?

Craig Barratt


From: Hiroshi Murakami <>
Date: Sun, 30 Jan 94 22:48:44 JST
Subject: Actual code for the Crawford's algorithm

Has anyone know from where the Crawford's algorithm
written in Algol-60 or Fortran or C is available ?

C.R.Crawford(1973). ``Reduction of a Band Symmetric Generalized
Eigenvalue Problem,'' Comm. Assoc. Comp. Mach. 16, 41-44.

I don't think this algorithm can be found in most standard math libraries.

Thank you.
Hiroshi Murakami


From: Alan Edelman <edelman@BOURBAKI.MIT.EDU>
Date: Mon, 24 Jan 94 21:40:58 EST
Subject: Large Linear Algebra Survey


Many of you may be familiar with my series of surveys designed to
assess users' needs and give perspective on the state of computing.
I have learned that such surveys are difficult because computational
scientists come from so many different fields. I annually read
many technical papers outside of my field of specialty -- each with
its own jargon. Though this is time consuming, I think the information
gathered has proved quite valuable.

My last survey may be found in the Journal of Supercomputer
Applications, 7 (1993), 113-128.

This year I would like to ask about both sparse and dense problems.
^^^^^^ ^^^^^

1) Has the world yet seen the solution of a 100,000 x 100,000
dense linear system?

2) Has the world yet seen the computation of some or all eigenvalues
or eigenvectors of a dense symmetric 30,000 x 30,000 matrix?

3) Has the world yet seen the computation of some or all eigenvalues
or eigenvectors of a dense non-symmetric 10,000 x 10,000 matrix?

A) I would like to hear about any solutions to an n x n system
of equations where n > 10,000 for dense problems, or
n > 100,000 for sparse problems. I would like to know
what is your solution method, where the problem arises,
and references to your papers and also an elementary paper
or text on the subject. Also, I would like to know
about the time and accuracy of your solution.

B) I would like to hear about eigenvalue problems, again
with n>10,000 for dense problems, or n>100,000 for
sparse problems. Please answer the same questions
as in A.

c) If you have a large problem, how would your work be
influenced if we could solve problems 100 times
as large in the same amount of time?

Thanks for your cooperation. This survey seems to have a wide
readership and your work will be duly noted. If you have participated
before, I look forward to your updates. If not, I look forward to
hearing from you.

Thanks again
Alan Edelman
Dept of Mathematics
MIT 2-380
Cambridge, MA 02139


From: Nick Higham <>
Date: Tue, 25 Jan 94 14:26:46 GMT
Subject: Cooley and Tukey FFT paper is a Citation Classic

NA Digest readers may be interested to know that a ``Citation Classic
commentary'' has been published by Cooley and Tukey describing the
background to their 1965 paper ``An Algorithm for the Machine
Calculation of Complex Fourier Series'' \cite{coto93}.
This paper, which describes the fast Fourier transform, is one of the
most-cited papers in mathematics and computer science, as measured by
the Science Citation Index. Current Contents is a TV Guide-sized
weekly listing of journal contents pages that is quite likely to be
available in your institution's library.
An earlier paper by Cooley describes the background to the FFT
algorithm in somewhat more detail \cite{cool90}.

author = "James W. Cooley and John W. Tukey",
title = "On the Origin and Publication of the {FFT} Paper",
journal = "Current Contents",
number = "51-52",
pages = "8-9",
year = 1993,
note = "Dec. 20-27",
mynote = "Citation Classic"

author = "James W. Cooley",
title = "How the {FFT} Gained Acceptance",
editor = "S. G. Nash",
booktitle = "A History of Scientific Computing",
publisher = "ACM Press, Addison-Wesley",
address = "Reading, Massachusetts",
pages = "133-140",
year = 1990

Nick Higham
Department of Mathematics
University of Manchester


From: Roger Haggard <>
Date: Tue, 25 Jan 1994 11:24:19 -0600 (CST)
Subject: Parallel Processing for Students

Can anyone give me a hint about getting access to some parallel
multi-processor machines for student use in a graduate course
on parallel computer architecture? If we could get some access
time over Internet to one or more machines, it would help the
students a lot. I've heard that NSF gives grants for this kind
of thing, but I don't remember any details.


Dr. Roger L. Haggard, Assistant Professor
EE Department, Box 5004
Tennessee Tech University Email: RLH0750@TNTECH.EDU
Cookeville, TN 38505 Phone: (615)372-3453


From: Walter Egli <>
Date: Fri, 28 Jan 94 09:42:35 +0100
Subject: Gradient of 3D Potential

Gradient of a potential in 3D.
I have a potential field phi in a n1*n2*n3 mesh.
I am looking for grad(phi) especially for good values at the
boundaries. Is there an easy procedure to do this ?
Walter Egli
ABB Corporate Reseacrh
CH-5406 Baden-Daettwil
Fax +41 56 83 45 69


From: Volker Mehrmann <volker.mehrmann@Mathematik.TU-Chemnitz.DE>
Date: Fri, 28 Jan 94 13:38:17 +0000 (MEZ)
Subject: Benchmarks for Systems and Control Theory

Second call for benchmark examples:

The Working Group on Software WGS, which consists of a number of
experts in control theory, numerical analysis and software design
from Belgian, Dutch, English and German Universities, Research Institutes
and Industry, develops in cooperation with the Numerical Algorithms Group NAG
from Oxford a Subroutine Library In COntrol Theory, called SLICOT.

The original goal of the SLICOT endeavor is the realization of a portable
library of basic, high-quality routines for systems and control theory and
signal processing, particularly with respect to numerical reliability
and computational efficiency. The current release 2.0 includes about 90
user callable FORTRAN 77 routines, mainly for the analysis, synthesis and design
of linear time-invariant control systems.

Since numerical robustness and efficiency are the basic requirements for all
contributions to future releases, the group is looking for good benchmark
examples for the evaluation and testing of numerical algorithms in this area.
We are particularly interested in examples which represent ill-conditioned
problems and examples which test the limits of the numerical methods with
respect to parameters like accuracy and speed.

Special areas of interest are the following:
Reduction to state space forms, minimal realization
Filtering problems
Riccati, Sylvester and Lyapunov equations
Pole placement problems
identification problems
variable coefficient problems
descriptor systems in control
robust control problems
adaptive control

Since our first call for benchmarks we have collected the following examples:

a) A test model data set for helicopter vibration:

Refer.: J. Petry, Zur Ansteuerung von Servoaktuatoren fuer die
Stoergroessenkompensation mittels Abtastsystemen.
DLR (Deutsche Gesellschaft fuer Luft und Raumfahrttechnik)
Report 86-08

J. Petry, Lineare Gleichungen fuer ASIS-Schwingungsmodell incl. Daten.
DLR (Deutsche Gesellschaft fuer Luft und Raumfahrttechnik)
Report ARNO-R 81-01

b) Identification:

Refer.: E.W. Jacobson and S. Skogestad, Identification of ill-conditioned
plants : a benchmark problem.

c) H-2, H-infinity control and model reduction:

Refer.: Y. Ge, L.T. Watson, E. Collins and D. Bernstein, Globally convergent
homotopy algorithms for the combined H-2/H-infinity Model reduction problem.
Report Virginia Polytechnical Institute

d) Descriptor Systems:

Refer.: B. Simeon, F. Grupp, C. Fuehrer, P. Rentrop, A nonlinear truck model
and its treatment as a multibody system.
Report TUM-M9204 Techn. Univ. Muenchen, 1992

T. Schmidt and M. Hou, Rollringgetriebe, Report Bergische Univ. Wuppertal 1992

e) Algebraic and differential algebraic Riccati equations:

Refer.: P. Benner, Ein orthogonal symplektischer Multishift Algorithmus zur
Loesung der algebraischen Riccatigleichung Diplomarbeit, RWTH Aachen, 1993

P. Kunkel and V. Mehrmann, Numerical Solution of Riccati Differential
Algebraic Equations.
Linear Algebra and its Applications, Vol. 137/138 , 1990 , pp. 39--66.

We would be very happy to receive further test problems and benchmark examples.

Examples in form of FORTRAN subroutines, MATLAB m-files or hardcopy papers
are welcome. They should contain values for all used parameters, as well as
a reference solution for which the accuracy is known.

The benchmark examples should be sent to

Volker Mehrmann Tel.: 049 (0)371-561-2659
Fachbereich Mathematik Fax.: 049 (0)371-561-2657
TU Chemnitz-Zwickau email:
PSF 964
D-09009 Chemnitz


From: John Brown <>
Date: Fri, 28 Jan 94 08:13:10 -0800
Subject: Fast Matrix Multiplication

Can anyone point me to some current references on the state-of-the-art
in fast matrix multiplication algorithms. The last survey I have is
from Victor Pan in 1988.

Please send to Thanks in advance.

John Brown


From: Jack Dongarra <>
Date: Fri, 28 Jan 94 12:49:55 -0500
Subject: Availability of ScaLapack

As part of the ScaLapack project, several new software items are now
available on netlib. Be aware that these are preliminary version
of the package. Major changes will occur over time. The new items
that have been introduced are:

1) Distributed memory version of the core routines from LAPACK
2) Fully parallel package to solve a symmetric positive definite sparse linear
system on a message passing multiprocessor using Cholesky factorization.
3) A package based on Arnoldi's method for solving large scale nonsymmetric,
symmetric, and generalized algebraic eigenvalue problems.
4) C version of LAPACK
5) LAPACK++ a C++ implementation of some of the LAPACK.
6) Templates for sparse iterative methods for non-symmetric Ax=b.

For more information on the availability of each of these
packages, consult the scalapack, clapack, c++, or linalg indexes on netlib.

echo "send index from scalapack" | mail
echo "send index from clapack" | mail
echo "send index from c++/lapack++" | mail
echo "send index from linalg" | mail

1) Distributed memory version of the core routines from LAPACK

Beta version 1.0 of this part of the package includes factorization
and solve routines for LU, QR, and Cholesky; decomposition routines to
Hessenberg form, tridiagonal form, and bidiagonal form; and,
preliminary versions of QR with column pivoting, triangular inversion,
and a parallel implementation of the SIGN function, which uses
deflation to calculate eigenvalues. Condition estimation and iterative
refinement routines are also provided for LU and Cholesky. The current
version of ScaLapack is in double precision real. Future releases of
ScaLapack will include complex versions of routines as well as the
single precision equivalents. At the present time, ScaLapack has been
ported to the Intel Gamma, Delta, and Paragon, Thinking Machines CM-5,
and PVM clusters. We are in the process of porting the BLACS to the

A second release of PUMMA (Parallel Universal Matrix Multiply Algorithm)
is included with the ScaLapack software. Both a PICL implementation and
a BLACS implementation of PUMMA are provided.

2) Fully parallel package to solve a symmetric positive definite sparse linear
system on a message passing multiprocessor using Cholesky factorization.

CAPSS (CArtesian Parallel Sparse Solver) is a fully parallel package to
solve a symmetric positive definite sparse linear system on a message
passing multiprocessor using Cholesky factorization. All phases of the
computation, from ordering through numerical solution, are performed in
parallel. The ordering uses Cartesian nested dissection based on an
embedding of the problem in Euclidean space. This first release is
meant for Intel iPSC/860 machines; the code has been compiled and
tested on an Intel iPSC/860 with 128 processors. The code is written
in C with message passing extensions provided by PICL (Portable
Instrumented Communications Library), which is also available from
netlib. CAPSS also uses a few native iPSC/860 functions.

3) A package based on Arnoldi's method for solving large scale nonsymmetric,
symmetric, and generalized algebraic eigenvalue problems.

ARPACK is a Fortran 77 software package for solving large scale
eigenvalue problems. The package is designed to compute a few
eigenvalues and corresponding eigenvectors of a large (sparse) matrix.
The package provides a communication interface(RCI) to user applications.
RCI allows maximal flexibility with respect to user needs and allows
(and requires) a user to define its own matrix-vector multiply
and/or linear solver routines for the ARPACK supported modes
(simple REGULAR, simple SHIFT-AND-INVERT, generalized REGULAR,
generalized SHIFT-AND-INVERT and CAYLEY mode are supported).
A symmetric ARPACK Intel Touchstone Delta parallel implementation
is also available on netlib (see arnoldi-delta/SRC/ex-sym.doc).
ARPACK depends on standard BLAS (Levels 1 , 2 and 3) and LAPACK libraries
which exist in object form on the Delta.

4) C version of LAPACK

CLAPACK is an automated f2c conversion of Fortran 77 LAPACK into ANSI C.
Be aware that since this is an f2c conversion of existing column-oriented
Fortran 77 LAPACK code, all CLAPACK code is column-oriented NOT

5) LAPACK++ a C++ implementation of some of the LAPACK.

LAPACK++ is the C++ version of LAPACK. This version includes support
for solving linear systems using LU, Cholesky, and QR matrix factorizations.
LAPACK++ supports various matrix classes for vectors, non-symmetric
matrices, SPD matrices, symmetric matrices, banded, triangular,
and tridiagonal matrices; however, Version 0.9 does not include all
of the capabilities of original f77 LAPACK. Emphasis is given to
routines for solving linear systems consisting of non-symmetric matrices,
symmetric positive definite systems, and solving linear least-square systems.
Support for eigenvalue problems and singular value decompositions are
not included in this prototype release. Future versions of LAPACK++
will support this as well as distributed matrix classes for parallel
computer architectures.

6) Templates for sparse iterative methods for non-symmetric Ax=b.
We have put together a book on iterative method for large sparse
nonsymmetric systems of linear equations. The book is available in
postscript form on netlib or can be ordered from SIAM.
Using concept of templates, we presents the algorithms using the
same notation in a straight forward manner permitting the user to
inspect, modify, or ignore any desired level of implementation detail.
Hints on parallelization, use, and other practical aspects are provided.

In addition to the algorithmic description in the book we have
provided a set of software in Fortran and in Matlab for the
following methods:

Bi-conjugate Gradient
Bi-conjugate Gradient stabilized
Conjugate Gradient
Conjugate Gradient squared
Generalized Minimal Residual
Quasi-Minimal residual
Successive Over-Relaxation

The ScaLapack group:
Oak Ridge National Laboratory
Rice University
University of California, Berkeley
University of Illinois
University of Tennessee

Comments and questions can be sent to


From: Keith Moore <>
Date: Fri, 28 Jan 1994 21:06:51 -0500
Subject: Administrivia: Changes to NA-Net Handles

There are now over 3600 members of NA-Net With this many members,
it is not surprising that there are several duplicated last names.

When you join NA-Net the NA-Net software assigns a "handle" or "key".
Until recently, the first person with a given last name would be
assigned that name as his or her key. The next person would get a
key of (first initial + last name), and so on.

We have had several complaints of mail being mis-routed because it
was being sent to, and winding up
somewhere else than the sender intended.

Therefore, all NA-Net members who had handles identical to their
last names, where there was someone else with the same last name,
have had their handles changed to make them unambiguous. In most
cases the handle is now (first initial + last name), but occasionally
it was necessary to construct a different name.

If you now send mail to one of these NA-Net addresses, you will
receive a message telling you that the address you supplied was
ambiguous, and giving you a list of addresses of people who
have that name.

You can also send mail to
If there are no ambiguities the mail will be forwarded to the
appropriate person.

Keith Moore

The following last names were affected:

addison dibenedetto jpetersen page sullivan
alexander douglas juang pan sun
anderson drake jumarhon papadopoulos sweet
arnold du kearsley park tam
bai duncan kelley patricio tam
baker edwards kim patterson tang
baldwin ellis kincaid paul taylor
barker elmroth klein pearson thomas
barrett emendes kreiss peng thompson
bartels eng krogh perry tsai
barton england kuo peters tsao
baum evans lai petersen turek
beckwith fdesprez lambert phillips turner
bennett fdibenedetto landman pierce van
berger feng lang plab varga
berman fernando lawson poliashenko vu
bharadwaj ferreira lee powell walker
biegler fortin li prevost wang
boman fox lin price ware
bradshaw french lindquist qin watson
brenner fulton lindsay randall weiss
briggs gaffney liu rao wheeler
brown gates lohner reid white
buratynski gilbert long reilly whitley
burke gill lou ren williams
byrne gong lu richardson wilson
cai goodman luk roberts wong
cameron gray ma robertson woo
carey green macdonald robey woo
carlson greif macleod robinson wu
carter gu mai rodrigues xie
chang guenette maier rosen xu
chen guo mann russo xue
cheney gupta mathew sand yan
cheng haggard mccarthy saunders yang
chou haggard mendes schmid ye
chow hamed meyer schneider yeh
chronopoulos hanke miller schultz young
ciarlet hari mitchell schulz yuan
climent hart mittal schwarz yun
cohen hasegawa mohanty segal zasadil
collins he moore shao zhan
cox henderson murphy shaw zhang
craig hermann murray shen zhao
crivelli higgins myers shi zheng
dai hill naik shyue zhou
david hoffman nash sigmon zhu
davies holmstrom navarra silva
davis hopkins nayar singer
day hu nelson smirnov
dekker hutchinson ng smith
delong hwang nielsen so
delves jennings oliveira soto
dempsey jensen olsen souli
deng johnson ostrouchov stevenson
desprez joubert owen stewart


From: Iain Duff <>
Date: Sat, 29 Jan 94 12:45:51 GMT
Subject: IMANA Newsletter Available via FTP

IMANA Newsletter Volume 18(2). January 1994.

The part of the January issue of the IMANA Newsletter that I have available
electronically can be accessed through anonymous ftp to RAL. The details
of how to access it are given below.

If readers wish to receive complete paper copies on a regular basis they
should write to:

Karen Jenkins
Catherine Richards House
Institute of Mathematics and its Applications
16 Nelson Street
Essex SS1 1EF

who will supply further information and subscription rates.

To get a copy


When prompted for a userid, reply with


and give your email address as a password.

Then directory to pub/open (cd pub/open)

Set mode to binary (bin)

Copy is in file imana.jan94.Z (get imana.jan94.Z)


From: Joseph Traub <>
Date: Tue, 25 Jan 94 07:37:26 EST
Subject: Joint Summer Research Conference


The Joint Summer Research Conferences are sponsored by the American
Mathematical Society, the Institute of Mathematical Statistics and the Society
for Industrial and Applied Mathematics. There will be a summer research
conference in continuous algorithms and complexity from Saturday June 11 to
Friday June 17, 1994 at Mount Holyoke College, Massachusetts.

There will be over thirty international invited speakers at the

Those interested in attending the conference should send the
following information to the Summer Research Conference Coordinator, Confer-
ences Department, American Mathematical Society, P. O. Box 6887, Providence,
RI 02940. Fax: 401-455-4004;e-mail:

1. Title and date of conference.
2. Full name.
3. Mailing address.
4. Telephone number and area code for office and home, e-mail address, fax
5. A short paragraph describing your scientific background relevant to the
topic of the conference.
6. Financial assistance requested; please estimate cost of travel.
7. Indicate if support is not required and if interested in attending even
if support is not offered.


The following description of the Conference is taken from the Notices of the

Saturday, June 11, to Friday, June 17
Continuous algorithms and complexity
J. F. Traub (Columbia University), co-chair
J. Renegar (Cornell University), co-chair

Continuous computational complexity studies the complexity of contin-
uous mathematical problems. The motivation is that most problems in the
physical, biological, and social sciences, and even in finance, have contin-
uous models.
The mathematical problems that arise in this modeling include integral
equations, multivariate integration, optimization, ordinary differential
equations,partial differential equations, and systems of polynomial equations.
Continuous computational complexity is applied in many disciplines
including scientific computation, control theory, and statistics. Invited
speakers will cover the foundations of continuous computational complexity,
recent results, open problems, and applications.
Continuous computational complexity may be contrasted with discrete
computational complexity, that is, the study of the complexity of problems
such as scheduling and networking. Much of the research in theoretical
computer science over the past twenty-five years has been devoted to discrete
Continuous computational complexity is a particularly good area for
researchers in the mathematical sciences. There is intense international
research activity. There are numerous important open problems which, while
deep, are solvable. Furthermore, progress requires mathematical tools such
as algebra, algebraic geometry, functional analysis, analysis in the large,
and measure theory.

Particularly promising new areas of computational complexity include:
Theoretical topics: Average case analysis of nonlinear problems, average
case complexity, complexity of multidimensional problems, relations between
condition numbers and complexity theory, models of parallel computation,
noisy data, power and limitations of randomization, and structural computa-
tional complexity.

Applications: Approximation, computer vision, dynamical systems, control
theory, ill-posed problems, integral equations, multivariate integration,
nonlinear equations, nonlinear optimization, ordinary differential equations,
partial differential equations, prediction and estimation, signal processing,
singularity detection, zeros of systems of polynomials, and systems of
polynomial equations and inequalities.

Members of the Organizing Committee: Lenore Blum (Mathematical
Sciences Research Institute and ICSI) and Henryk Wozniakowski (University
of Warsaw and Columbia University).


From: Gene Golub <golub@sccm.Stanford.EDU>
Date: Fri, 28 Jan 94 13:34:25 PST
Subject: NA Summer School -- 1995

Mathematics of Numerical Analysis
Real Number Algorithms
for a Four-Week Conference for Summer 1995
Steve Smale (

The mathematical theory of real number algoritms is to be the
subject of this proposed conference. Thus numerical analysis will be
central with emphasis on geometrical, algebraic, analytic and foundational
perspectives. Investigations of effciency will play a special role.
Pratical algorithms will be the subject of theoretical analysis, but immediate
useful results will not be demanded.
It is to be hoped that conference will give the subject of numerical
analysis a greater coherence through a focus on the mathematical side.
In particular, an aim to strenghten the unity of mathematics and numerical
analysis, and to narrow the gap between pure and applied mathematics. That
goal is appropriate since many of the heroes of pure and applied mathematics,
Newton, Euler, Lagrange and Gauss among them, established the basic real
number algorithms. With the revolution of the computer and the great
achievements of scientific computation, it does service to both the pure and
applied communities to support the mathematical development of numerical
analysis .
This is an appropriate time to schedule such a meeting in view of the
rapid development of heuristic work, a good base of theoretical work, and
a widespread desire for mathematical deepening of the subject.
The starting date is July 16, 1995 and is to last 4 weeks. The
place is to be Park City, Utah. Appropriate reservations have been made by
the American Math. Society.
The conference would be international in character with strong
representation form the most mathematically developed parts of numerical
analysis. Besides tutorials and short courses, seminars in the following
areas are contemplated:

(a) Linear Algebra
(b) Non-linear systems-path following
(c) Differential equations
(d) Linear programming problems
(e) Algebraic questions
(f) Foundations
(g) Information based complexity
(h) Lower Bounds
(i) Approximation theory

Organizing Committee for
Mathematics of Numerical Analysis
Park City, July 16, 1995, for 4 Weeks

Steve Smale, Chair, University of California, Berkeley
Gene Allgower, Colorado State University
Lenore Blum, MSRI, Berkeley
Alexander Chorin, Berkeley
Philippe Ciarlet, Universite Pierre et Marie Curie, Paris
Felipe Cucker, Universitat Pompeu Fabra, Barcelona
James Demmel, Berkeley
Ron DeVore, University of South Carolina
Gene Golub, Stanford
Arieh Iserles, University of Cambridge, England
Bert Jongen, Aachen, Germany
Herb Keller, Cal Tech
Jim Renegar, Cornell
Mike Shub, IBM, Yorktown Heights
Gill Strang, MIT
Henryk Wozniakowski, University of Warsaw and Columbia University


Date: Sun, 30 Jan 94 14:05:40 IST
Subject: Workshop on Total Positivity

An International Workshop On
Total Positivity and its Applications
Jaca, Spain, September 26--30, 1994

Organised by Departamento de Matematica Aplicada
University of Zaragoza, Spain

Total Positivity has proved to be a powerful tool in many areas of
pure and applied mathematics. However there has never been a meeting
which focused solely on this subject. In view of the diverse
applications of total positivity we feel that a meeting which
draws together experts whose lectures will cover the principal areas of
applications and participants whose research can benefit from these
techniques will be a timely and important scientific event.
Among the topics which we plan to have represented at the conference
are the applications of total positivity to Probability and Statistics,
Combinatorics, Integral and Differential Equations, Geometric Modeling,
Matrix Theory, Approximation Theory, Complex Analysis, Numerical Analysis
and Wavelets Analysis. In each of these fields there are problem
areas in which total positivity has proved to be useful, if not
indispensable. We intend to bring together as many researchers
as possible who use total positivity in their work or are
interested in learning about the latest developments in this subject.

Organising Committee
Prof. Mariano Gasca
Dr. Charles A. Micchelli
Prof. Allan Pinkus
Dr. Timothy N.T. Goodman

Partial List of Invited Speakers

B. Bojanov (Sofia), F. Brenti (Perugia), J. Carnicer (Zaragoza),
J. Garloff (Konstanz), M. Gasca (Zaragoza), T.N.T. Goodman (Dundee),
B. Heiligers (Augsburg), R-Q. Jia (Edmonton), S. Karlin (Stanford),
K. Morken (Oslo), J. M. Pena (Zaragoza), A. Pinkus (Haifa),
H. Pottmann (Vienna), R. Zalik (Auburn).

Local Organising Committee

J.M.Carnicer, M. Garcia, M.C. Lopez de Silanes, J.J.Martinez, J.M.Pena.

Correspondence and Further Information

Those interested in participating, contributing or receiving future
announcements, please contact the organisers:

IWTPA, Depto. Matematica Aplicada
Facultad de Ciencias
Edificio de Matematicas
Universidad de Zaragoza
50009 Zaragoza, Spain
Fax: (34)76 356244 Phone: (34)76 356617
E-mail: or


From: Kaj Holmberg <>
Date: Tue, 25 Jan 94 14:37:09 +0100
Subject: Position at Linkoping University, Sweden

Please circulate this to everyone who might be interested.


Full Professor of Operations Research

Linkoping University seeks to fill the chair in Operations Research/
Mathematical programming. The successful candidate will be expected
to provide leadership to a growing Optimization group whose interests
span a wide range of operations research areas with special emphasis
on mathematical programming. The Optimization group includes six
faculty members. The position involves teaching at both the under-
graduate and graduate levels, including the supervision of Ph.D.
students, in addition to engaging in and running a number of active
research programs.

For more information contact Rector Professor Sven Erlander,
+46(13)281001, or Asst Professor Jan Lundgren, +46(13)281409

Persons interested should send a brief resume as soon as possible.
Further details should be sent on request. Send to:

Department of Mathematics
Division of Optimization
Institute of Technology

Telefax: +46(13)100746


From: Steve Pruess <>
Date: Tue, 25 Jan 1994 09:39:07 -0700
Subject: Department Head Opening, Colorado School of Mines

Colorado School of Mines

The Colorado School of Mines is seeking candidates for the position of
Head of the Department of Mathematical and Computer Sciences. This
department offers BS, MS, and PhD degrees under the department title.
With a faculty of 18 tenured and tenure track members, the department
annually receives roughly a million dollars in grants; 116 undergraduate
students and 70 graduate students are currently enrolled in our
degree programs.

The position requires a PhD in a mathematical or computer science.
The applicant should have a sufficiently outstanding record of scholarly
achievement and teaching experience to justify a tenured appointment at
the Full Professor level. In addition, the successful applicant must
have held an academic position for at least five years, and show evidence
of demonstrable administrative ability, including visionary leadership,
communication skills, and effective interaction and evaluation of personnel.
The Head is expected to manage and direct the department's efforts in
instruction and in scholarship, to continue and enhance its excellence in
teaching and in research, to plan and oversee the development of its
research activities and academic programs, and to represent the department
on campus and externally.

The Colorado School of Mines is a state university, internationally renowned
in the energy, materials, and resource fields, attracting outstanding
students in a broad range of science and engineering disciplines. The School
of Mines is strongly committed to quality teaching and research. CSM provides
an attractive campus environment, a collegial atmosphere, relatively small
size (3000 students, about 30% in graduate programs), and an ideal location
in the foothills of the Rocky Mountains 13 miles from downtown Denver.

Applications will be considered beginning February 15, 1994 and thereafter
until the position is filled. The applicant should provide a statement giving
administrative, pedagogical, and scholarly philosophy which should include
a discussion of advantages and disadvantages of programs combining both
Computer Science and Mathematics, and how to reconcile research vs. teaching
conflicts. This letter and a vita should be sent by postal mail to

Colorado School of Mines
Department Head Search #94-01-31
1500 Illinois Street
Golden, CO 80401

The applicant must also arrange for five letters of reference to be mailed
to the above address or sent by email to

CSM is an Affirmative Action/Equal Opportunity Employer. Women and
minorities are encouraged to apply.


From: Daniel Okunbor <>
Date: Tue, 25 Jan 94 13:27:34 CST
Subject: Chair of Computer Science at University of Missouri-Rolla

University of Missouri-Rolla

The Department of Computer Science at the University of Missouri - Rolla
is seeking a dynamic individual for the position of Chairperson. Applicants
must have an established record of scholarship, research, and leadership
sufficient to justify a tenured position as Professor of Computer Science.
The Department is seeking a person who will be a strong proponent for both the
graduate and undergraduate research efforts and degree programs in
computer science.

The Department grants the B.S., M.S. and Ph.D. degrees. The Ph.D.
program has been active since 1977 and the Department currently
has over 100 graduate students. Departmental research is growing
with current funding above a million dollars from NSF, DoD, and industry
sources. Major computing facilities include an Intel iPSC/860 32 processor
multicomputer as well as SUN, SGI, and NeXT workstations for faculty and
student use. Disciplinary and interdisciplinary research within the Sciences
and Engineering is active in Parallel and Distributed Computing, Scientific
Computing, Formal Methods, and Artificial Intelligence.

The University of Missouri-Rolla is the primary science and engineering
campus of the University of Missouri system; it currently has an enrollment
of 5200 students. Rolla is situated in the non-urban environment of
the Ozarks equidistant from St. Louis, Columbia, and Springfield.

The committee will begin reviewing applications on February 21, 1993.
Applicants should send a complete resume, including the names, addresses,
and phone numbers of at least four references to:

Chairperson Search Committee,
Department of Computer Science,
University of Missouri-Rolla, Rolla, MO 65401,
(314)-341-4491 (

The search committee also solicits nominations of qualified candidates for the
position. UM-Rolla is an equal opportunity/affirmative action employer
and strongly encourages applications from women and members of minority groups.


From: Tom Hayden <>
Date: Tue, 25 Jan 1994 14:52:14 EST
Subject: Post-doctoral Position at University of Kentucky

University of Kentucky--Department of Mathematics

The department invites applications for a post-doctoral associate
position for two years contingent on NSF funds from a continuing
grant that extends to July 1996. The research is in the application
of distance geometry to molecular conformations. The applicant is
expected to continue research on algorithms and produce user friendly
software for chemists/biochemists. The applicant should have a background
in developing software using C or C++, be able to understand numerical
linear algebra/nonlinear optimization, and have a chemistry background
at least through organic chemistry (better biochemistry).

Applicants should send a resume specifically mentioning expertize in the
above areas and at least two letters of reference to: T. L. Hayden,
Department of Mathematics, University of Kentucky, Lexington Kentucky,
40506. The University of Kentucky is an affirmative action/equal
opportunity institution.


From: M Berzins <>
Date: Thu, 27 Jan 94 19:09:13 GMT
Subject: Postion at University of Leeds

University of Leeds, School of Computer Studies
Parallel Mesh Generation and Adaptivity.

A three year SERC funded research fellowship is available at
Leeds as part of a joint project with the University College
of Wales at Swansea. The aim of the project is to design
and write parallel meshing and adaptivity software for
unstructured tetrahedral meshes. The main Leeds and Swansea
supervisors are Dr. M. Berzins and Dr N.P. Weatherill.
The Salary is on standard UK academic scales and the post is available
from 1st March or as soon as possible thereafter.
The post will involve working with oil and aerospace companies.
Enquiries by email to .


From: Maui HPCC <>
Date: Thu, 27 Jan 1994 15:36:42 -1000
Subject: Position at Maui High Performance Computing Center


Application Period: 1/15/94 to 2/15/94

SUMMARY OF DUTIES: Research Scientist/Engineers will be responsible
for a range of user services activities including customer training and on-going
customer support. This may include developing center tools and application
software, consulting with customers on parallization strategies for their
applications, developing training materials, teaching customer education
classes, supporting the customer help desk, and maintaining an extensive on-line
customer help system. The computing equipment at the Maui center is currently
a 32-node IBM SP1. By the end of 1994, equipment will include a 400-node
IBM SP2 and an 80-node IBM SP2. The latter will be dedicated to Department
of Defense classified work. Additionally, the Center has SGI and other
visualization computers and a number of workstations.

A Bachelor's in Engineering, Science, or Mathematics and two years of
experience or an MS with one year experience is required. Advanced course
work and experience may be substituted on a one year for one year basis for
some educational requirements.

A Bachelor's in Engineering, Science, or Mathematics and three years of
experience or an MS with two years experience, or a Ph.D. is required.
Advanced course work and experience may be substituted on a one year for
one year basis for some educational requirements.

should have experience in scientific programming in a UNIX environment. An
in-depth knowledge of FORTRAN and/or C is highly desirable. Experience
with distributed memory parallel programming is desired. Candidate should be
self-motivated and have good interpersonal skills. Applicants with a background
in scientific or engineering discipline with computing experience are encouraged
to apply.

CONDITIONS OF EMPLOYMENT: Must have or be able to obtain an Air
Force top secret clearance. Successful candidate will be required to provide
official college transcripts, certifications and licenses that may be required
for this position. Must be able to work flexible hours and attend community
or governmental meetings.

TO APPLY: Applications/resumes must be received by Human Resources at
1717 Roma NE, Albuquerque, New Mexico 87131 or Maui High Performance
Computing Center, 535 Lipoa Parkway, #201, Kihei, Maui, HI 96753 no later
than 5:00PM on the closing date. Resumes must be accompanied by a cover
letter with an original signature. Indicate the job title on application/cover

The University of New Mexico is an Affirmative Action/Equal Opportunity
Employer and Educator


From: Iain Duff <>
Date: Fri, 28 Jan 94 18:19:37 GMT
Subject: Position at CERFACS

Immediate employment prospect. CERFACS, Toulouse, France

If you fancy a year or two in SW France starting as soon as you can book a
flight, the following might be of interest.

We are looking for someone at post-doc/senior level with experience, interest,
and background (with publications) in iterative methods for solving
linear systems, preferably unsymmetric systems.

We have two possible positions. One that is open to any nationality and is
covered by an ESPRIT contract. For this we would like someone with experience
both on distributed computers and on PDEs solutions. Basically we have to
support an effort for the parallelization of a multi-block explicit CFD solver
on distributed computers (using message passing). The intention is that
about 50% of the time of this person would be devoted to research on
iterative methods.

The second position is an HCM grant and so is restricted to non-French
European nationals. The primary focus of this HCM research network will
be the study of iterative methods and preconditioners for the solution of
linear systems. Special emphasis will be placed on developing techniques
that are well-suited to highly nonsymmetric and/or indefinite
systems arising from computational fluid dynamics and electromagnetics.
Our partners in this European project include: CNR, CRS4, Utrecht, CEA and

Please send email to


From: John Guckenheimer <>
Date: Fri, 28 Jan 1994 18:06:08 -0500
Subject: Postdoc Position at Cornell

Postdoctoral Position
Bifurcation Analysis of Dynamical Systems

A postdoctoral position is available at Cornell University in the
Center for Applied Mathematics. The position is for research on the
numerical analysis of bifurcations in nonlinear systems of differential
equations. Requirements are a strong background in the mathematical
theory of dynamical systems and significant computing experience.

Applicants should send a vita, (p)reprints and have three letters of
reference to be sent to

John Guckenheimer, Director
Center for Applied Mathematics
Cornell University
Ithaca, NY 14853

Funding for this position is provided by a grant from the
U.S. Department of Energy


From: Ron Boisvert <>
Date: Mon, 24 Jan 94 11:58:04 EST
Subject: Contents, ACM Transactions on Mathematical Software

Table of Contents
ACM Transactions on Mathematical Software
Volume 19 Number 4 (December 1993)

Algorithm 722: Functions to Support the IEEE Standard for Binary
Floating-Point Arithmetic
W.J. Cody and Jerome T. Coonen
pp. 443-451

Algorithm 723: Fresnel Integrals
W. Van Snyder
pp. 452-456

Toward Parallel Mathematical Software for Elliptic Partial
Differential Equations
Calvin J. Ribbens, Layne T. Watson and Colin Desa
pp. 457-473

Applying Series Expansion to the Inverse Beta Distribution to Find
Percentiles of the F-Distribution
Roger W. Abernathy and Robert P. Smith
pp. 474-480

Algorithm 724: Program to Calculate F-Percentiles
Roger W. Abernathy and Robert P. Smith
pp. 481-483

A Remark on Algorithm 643: FEXACT: An Algorithm for Performing
Fisher's Exact Test in rxc Contingency Tables
Douglas B. Clarkson, Yuan-an Fan and Harry Joe
pp. 484-488

A Portable Random Number Generator Well Suited for the Rejection
W. Hormann and G. Deflinger
pp. 489-495

Rounding Errors in Certain Algorithms Involving Markov Chains
Winifred K. Grassmann
pp. 496-508

A Test Problem Generator for the Steiner Problem in Graphs
B.N. Khoury, P.M. Paradalos and D.-Z Du
pp. 509-522

Implementation of a Lattice Method for Numerical Multiple Integration
Stephen Joe and Ian H. Sloan
pp. 523-545

Corrigendum: Algorithm 725: Computation of the Multivariate Normal
Zvi Dresner
p. 546

1993 Author Index
pp. 547-549


From: Daniel Baltzer <>
Date: Thu, 27 Jan 1994 10:36:43 +0100
Subject: Contents, Advances in Computational Mathematics

Advances in Computational Mathematics, Volume 2, No.1, 1994, ISSN 1019 7168
Editors-in-Chief: John C. Mason & Charles A. Micchelli

special issue: Volume 2, No. 1, 1994: GEOMETRIC MODELING
Editor: Philip Barry

pp 1-21: NURBS approximation of surface / surface intersection curves
C.L. Bajaj and G. Xu

pp 23-40: Elastic curves on the sphere
G. Brunnett and P.E. Crouch

pp 41-66: Pythagorean-hodograph space curves
R.T. Farouki and T. Sakkalis

pp 67-80: A weak condition for the convexity of tensor-product Bezier and
B-spline surfaces
M.S. Floater

pp 81-99 Least squares surface approximation to scattered data using
multiquadratic functions
R. Franke, H. Hagen and G.M. Nielson

pp 101-122 Triangular finite elements of HCT type and class Cr
M. Laghchim-Lahlou and P. Sablonniere

pp 123-142 Helix splines as an example of affine Tchebycheffian splines
H. Pottmann and M.G. Wagner

pp 143-154 Convergence of subdivision and degree elevation
H. Prautzsch and L. Kobbelt

Submissions of articles and proposals for special issues are to be
addressed to the Editors-in-Chief:

John C. Mason
Applied & Computational Mathematics Group
Royal Military College of Science (Cranfield University)
Shrivenham, Swindon, SN6 8LA, England


Charles A. Micchelli
Mathematical Sciences Department
IBM Research Center
P.O. Box 218, Yorktown Heights, NY 10598, USA
E-mail: cam@yktvmz.bitnet

Free specimen copies and orders for Advances in Computational Mathematics
are to be sent to:

J.C. Baltzer AG, Science Publishers
Asterweg 1A
1031 HL Amsterdam
The Netherlands
tel. +31-20-637 0061
fax. +31-20-632 3651


End of NA Digest