NA Digest Sunday, March 7, 1993 Volume 93 : Issue 10

Today's Editor:

Cleve Moler
The MathWorks, Inc.

Submissions for NA Digest:

Mail to

Information about NA-NET:

Mail to


From: Stanly Steinberg <>
Date: Mon, 1 Mar 93 10:52:28 MST
Subject: Discontinuous Coefficients and Convergence Rates

Suppose L_h is a second-order finite-differrence or finite-element
approximation of

Lf = div A grad f

on a regular grid in n dimensions and where A is a smooth symmetric
positive-definate matrix. Now consider the case when A is not necessairly
continuous. I believe the correct assumption is that A is of bounded
variation. Numerical results indicate that L_h is now a ***first***-order
approximation of L. Some results of Babuska in one-dimension show that
certain finite element approximations of L are first-order (with no
restriction on the grid). The QUESTION is: Are there any results
for finite-difference methods that show first-order approximation.

More precisely if L_h f_h = g_h is a second-order approximation of
L f = g on a regular grid, then our numerical results show that the
sup norm error of (f_h-f) and (A grad(f_h-f)) are dominated by a C*h where
C is a constant.

Stanly Steinberg, Professor of Mathematics
Department of Mathematics and Statistics
Senior Research Scientist
Ecodynamics Research Associates, Inc.


From: David Wright <>
Date: Mon, 1 Mar 93 22:11:30 GMT
Subject: Incomplete Beta Function Derivatives

I would be grateful for any information on software or algorithms for
calculating 1st and 2nd order partial derivatives of the incomplete beta
function with respect to the two parameters (a and b) occurring in the
integrand of the standard integral form.

This would be helpful in maximum-likelihood fitting of censored or truncated
data in the cases of several parametric probability distributions.

Thanks in advance for any help,

David Wright

Mr. D R Wright, Tel. Voice: +44 71 477 8424
Center for Software Reliability, Fax: +44 71 477 8585
City University,
Northampton Square, e-mail:


From: Christian Schulz <>
Date: Tue, 02 Mar 93 11:29:56 SET
Subject: Examples for Overdetermined Systems

In connection with my thesis I'm working on solving systems of multi-
variate polynomial equations, especially overdetermined ones. Now I'm
looking for concrete examples of such systems arising in practice,
esp. problems in chemistry. Can anybody help me or give any hints where
I could look somewhere else? Thanks a lot, and have a nice time|
Christian Schulz < >


From: H. Murakami <>
Date: Thu, 4 Mar 93 06:32:00 JST
Subject: What is Best Method for Ax=lambda Bx ?

What is the currently known best methods for solving
the symmetric generalized eigenvalue problems
of LARGE size ?
Hiroshi Murakami


From: Daniel B. Szyld <>
Date: Wed, 3 Mar 93 18:02:59 EST
Subject: Change of Phone Numbers at Temple

Temple University is changing phone numbers.
The new numbers are:

David Hill (215) 204 1654
Daniel B Szyld (215) 204 7288
Department of Mathematics Fax (215) 204 6433


From: Alan Karp <>
Date: Mon, 1 Mar 93 08:29:00 -0800
Subject: Gordon Bell Prize for 1993

The 1993 Gordon Bell Prizes

The Gordon Bell Prizes recognize achievements in large-scale
scientific computing. Entries for the next Prize are due on 1 May
1993, and finalists will be announced by 30 June 1993. Pending
approval by the Supercomputing '93 program committee, finalists will
be invited to present their work at a special session of that meeting
in November 1993. Winners and honorable mentions will be announced
following the presentations.

The 1993 prizes will be given in two of three categories:

1. Performance: The entrant will be expected to convince the judges
that the submitted program is running faster than any other
comparable engineering or scientific application. Suitable evidence
will be the megaflop rate based on actual operation counts or the
solution of the same problem with a properly tuned code on a machine
of known performance, such as a Cray Y-MP. If neither of these
measurements can be made, the submitter should document the
performance claims as well as possible.

2. Price/performance: The entrant must show that the performance of
the application divided by the list price of the smallest system
needed to achieve the reported performance is better than that of
any other entry. Performance measurements will be evaluated as for
the performance prize. Only the cost of the CPUs, memory, and any
peripherals critical to the application need be included in the
price. For example, if the job can be run on diskless compute
servers, the cost of disks, keyboards, and displays need not be

3. Compiler parallelization: The combination of compiler and
application that generates the most speed-up will be the winner.
Speed-up will be measured by dividing the wall clock time of the
parallel run by that of a good serial implementation of the same
job. These may be the same program if the entrant can convince the
judges that the serial code is a good choice for a uniprocessor.
Compiler directives and new languages are permitted. However, anyone
submitting an entry in other than a standard, sequential language
will have to convince the judges that the parallelism was detected
by the compiler, not by the programmer.

There are some general conditions:

1. The submitted program must have utility; it must solve a problem
that is considered a routine production run, such as making daily
weather predictions or solving an important engineering or
scientific problem. It should not be a contrived or experimental
problem that is intended just to show high speed-up.

2. Entrants in the price/performance category must demonstrate that
the machine they used has real utility. (No fair picking up a few
used Z-80s for $1 each.) Only list prices of components should be
used. If the machine is not on the market, the entry is probably not
eligible although the judges will consider any reasonable estimate
of the price.

3. One criterion the judges will use for all categories is how much
the entry advances the state of the art of some field. For example,
an entry that runs at 15 Gflops but solves a problem in a day that
previously took a year might win over an entry that runs at 20
Gflops solving a more mundane problem. Entrants who believe their
submission meets this criterion are advised to document their claims

4. In all cases the burden of proof is on the contestants. The
judges will make an honest effort to compare the results of
different programs solving different problems running on different
machines, but they will depend primarily on the submitted material.

Contestants should send a three or four page executive summary to
Marilyn Potes, IEEE Computer Society, 10662 Los Vaqueros Circle, Los
Alamitos, CA 90720-2578 before 1 May 1993.


From: Melvyn Ciment <>
Date: Tue, 2 Mar 1993 09:34:25 -0500
Subject: NSF High Performance Computing

Dear Colleague,

Below is an extract of an announcement that will appear in the Federal Register
regarding the second meeting of the Blue Ribbon Panel on High Performance
Computing. This Panel was established at the direction of the National
Science Foundation's National Science Board.

Let me draw your attention to the set of four questions {Section 1) a)-d)}
contained in the Announcement. This set was extracted from questions and
issues discussed at the first meeting of the HPC Blue Ribbon Panel with the
intent of asking the scientific and engineering community to provide
comments to the panel.

Please feel free to share these questions with your colleagues or other
interested parties. Your comments are solicited; please do so in
accordance with instructions below.


Melvyn Ciment


DATE AND TIME: March 11, 1993: 8:30 a.m. to 5:30 p.m.
March 12, 1993: 8:30 a.m. to 2:30 p.m.

PLACE: Hotel Washington, 15th & Pennsylvania Ave., NW, Washington, DC 20004

To assess current knowledge on the state of supercomputing and the
contributions of high performance computing to scientific research and
education; to help project changes required by developments in this rapidly
evolving field; and to provide recommendations to the National Science
Board on NSF's possible future role in supercomputing.

AGENDA: 1) Invited panelists of computer designers and industry vendors,
NSF supercomputer center directors, and other experts will focus on the
following questions:

a) How would you project the emerging high performance computing
environment and market forces over the next five years and the implications
for change in the way scientists and engineers will conduct R&D, design and
productions modelling?

b) What do you see as the largest barriers to the effective use of these
emergent technologies by scientists and engineers and what efforts will be
needed to remove these barriers? What is the proper role of government, and,
in particular, the NSF to foster progress?

c) To what extent do you believe there is a future role for government-
supported supercomputer centers? What role should NSF play in this spectrum
of capabilities?

d) To what extent should NSF use its resources to encourage use of high
performance computing in commercial industrial applications through
collaboration between high performance computing centers, academic users and
industrial groups?

The panel welcomes comments from all interested parties. Persons wishing to
file written comments should mail (hard copy or electronic) or fax the
comments by April 1, 1993, to the Blue Ribbon Panel on High Performance
Computing, Room 306, National Science Foundation, 1800 G St., NW,
Washington, DC 20550. FAX # 202-357-0320. E-mail:
(Internet) or hpcmail@nsf (Bitnet).


From: Kevin Burrage <>
Date: Mon, 1 Mar 93 07:49:02 EST
Subject: Position at University of Queensland

The University of Queensland
Department of Mathematics

Applications are invited from researchers in any branch
of computational mathematics who wish
to spend a period of some weeks, up to a maximum of six months, in the
Department of Mathematics in 1994.

The computational group consists of

Professor Kevin Burrage : parallel computing, spatial modelling, odes
A/Professor Holt : operations research, scheduling algorithms
Dr Belward : integral equations, spatial modelling
Dr Chandler : integral equations, boundary elements
Dr Forbes : cfd, nonlinear phenomena
Dr Watts : cfd, operations research

The Fellowship may provide a stipend and a single economy return
airfare to Brisbane. The fellowships are available every year.
Application forms are available from

Dr. V.G. Hart
Department of Mathematics
The University of Queensland
Brisbane 4072

E-mail address:
Fax no. (07) 870 2272

Professor Kevin Burrage
email :


From: E. Wegman <>
Date: Sun, 28 Feb 93 18:29:38 -0500
Subject: Interface Symposium: Computing Science and Statistics

25th Symposium on the Interface: Computing Science and Statistics

Theme: Statistical Applications and Expanding Computer Capabilities
Date: April 14-17, 1993
Place: San Diego, California, Pan Pacific Hotel

Keynote Speaker: David Brillinger, "Statistics and Computing in Science"

Sponsor: Interface Foundation of North America

Cooperating Societies and Institutions:
American Statistical Association (ASA)
Institute of Mathematical Statistics (IMS)
Society for Indusrial & Applied Mathematics (SIAM)
Operations Research Society of America (ORSA)
The Biometrics Society (WNAR)
University of California, Berkeley
San Diego State University, San Diego
Northern and Southern California Chapters, American Statistical Association

Invited Sessions Include:
Data Compression
Computing with Enviromental Data
Biopharmaceutical Maps and Graphics
Clinical Trials
Protein Structure
Digital Networks
User Interfaces
Software Engineering and Statistical Methods
The Interface at 25
Likelihood Applications
Library Systems
Medical Applications
Multivariate Function Estimation
Networked Information Systems
Time Series Analysis
Computers and Statistics in Drug Discovery
Quality Data Bases

Enquiries should be sent to:
Interface '93
Michael E. Tarter, Program Chair
140 Warren Hall
University of California, Berkeley
Berkeley, CA 94720
(510) 642-4601
email: tarter@stat.Berkeley.EDU

Conference proceedings of invited and contributed papers will be published.
Camera-ready copy will be due on June 1, 1993.

Conference Schedule:
The conference begins on Wednesday evening, April 14, with a get-acquainted
reception. Technical sessions will be held Thursday and Friday with a banquet
Thursday evening. There will be final technical sessions Saturday morning.

The registration fee is $185 for members of the cooperating societies, ASA,
IMS, SIAM, ORSA, the Biometrics Society (ENAR and WNAR). The fee is $75 for
students. Please make checks payable to Interface '93.

General Information: THE PAN PACIFIC HOTEL, 402 West Broadway, San Diego,
CA 92102-3580. Telephone (619)239-4500 or (800)626-3988.


From: Ed Deprettere <>
Date: Mon, 1 Mar 93 10:57:20 +0100
Subject: Workshop on VLSI Signal Processing



An activity of the IEEE SP Society's Technical Committee on VLSI
organised in cooperation with IEEE Benelux, IEEE Benelux Chapter
on Signal Processing and EURASIP

October 20-22, 1993
Koningshof, Veldhoven, The Netherlands


The objective of this Workshop is to provide a forum for discussion of new
theoretical and applied developments in signal processing in its relation to
implementation as Very Large Scale Integrated Circuits.
The key note address and one panel session will focus on
"Industrial and Technical Challenges in Signal Processing for Consumer
Applications". The aim is to have also one session devoted to this topic.
A hard-bound record of the Workshop will be published.

Papers are solicited that relate to the technologies involved in the design
and implementation of signal processing algorithms and systems as VLSI

Digital Signal Processing
Algorithms Architectures
Languages Transformational design
Signal Processing Applications
Speech and music
Digital audio
Image and Video/HDTV
Computer Graphics
Radar and Sonar
Integrated Circuits and Systems
This area refers to the different phases, methods and tools (CAD)
used in designing signal processing algorithms and systems that may
lead to a final implementation in silicon:
Specification Design descriptions
Data and Control Flow
Design methodologies
Cell/Silicon compilers
Hardware/software co-design
Formal proofing
Testability and testing

Being organized for the first time in Europe presents a unique opportunity
to demonstrate chip sets developed in the JESSI program and to discuss
and demonstrate design methodologies that originated in the ESPRIT program.

Prospective authors are invited to submit before April 1, 1993 seven (7)
copies of a complete paper and an abstract for review category
classification to:

Mrs. M. Emmers/Mrs. M. van Kessel,
Philips International B.V./CPDC, Building VO-p, P.O. Box 218,
5600 MD Eindhoven, The Netherlands

GENERAL CHAIR: Ludwig Eggermont,
tel. +31(40)78 49 61, fax +31(40)78 64 22
Patrick Dewilde,
Ed Deprettere,
tel. +31(15)78 62 89, fax +31(15)62 32 71
Jef van Meerbergen,
Stefaan Note,
Takao Nishitani,
U.S. LIAISON: Bob Owen,

The Workshop will be held in Conference Centre "Koningshof", Veldhoven,
The Netherlands. Veldhoven is a 1.5 hour drive from the international
airports of Amsterdam, Brussels and Dussel-dorf and 10 minutes from
Eindhoven Airport.


April 1, 1993 Submission of paper and abstract
June 5, 1993 Notification of authors
July 15, 1993 Receipt of final photo-ready paper


From: Zhijian Huang <huang@ibguniv.bitnet>
Date: 01 Mar 1993 17:45:03 -0500 (EST)
Subject: NATO Advanced Study Institutes Programme

Algorithms for Continuous Optimization: The State of the Art
September 5-18, 1993
Il Ciocco, Castelvecchio Pascoli, Tuscany, Italy

COURSE DIRECTOR: Prof. Emilio Spedicato

Department of Mathematics
University of Bergamo
Piazza Rosate 2
24100 Bergamo
Tel. +39-35-277514,
Fax: +39-35-234693

SCIENTIFIC COMMITTEE: Prof. L.Dixon (Hatfield)
Prof. D.Shanno (Rutgers)
Prof. E.Spedicato (Bergamo)


F.Giannessi (Pisa): Optimality conditions
C.G.Broyden (Bologna): Numerical methods for linear systems
J.Martinez (Campinas): Algorithms for nonlinear algebraic equations
A.Bjorck (Linkoping): Algorithms for linear least squares
R.Fletcher (Dundee): Algorithms for unconstrained optimization
M.Biggs (Hatfield): Algorithms for constrained optimization
A.Conn (New York): Methods for large scale constrained optimization
G.Di Pillo (Rome): Exact penalty methods
J.Stoer (Wurzburg): Interior point methods for nonlinear optimization
E.Spedicato (Bergamo): Nonlinear optimization via ABS methods
R.Schnabel (Boulder): Parallel algorithms for nonlinear optimization
J.Zowe (Bayreuth): Nondifferentiable optimization
L.Dixon (Hatfield): Automatic differentiation and neutal network
J.Judice (Coimbra): Algorithms for complementarity problems
D.Shanno (Rutgers): Algorithms for linear programming
N.Deng (Beijing): Nonquadratic models in unconstrained optimization
/ Nonlinear programming in China
Y.Evtushenko (Moscow): Deterministic global optimization
/ Nonlinear programming in former Soviet Union


Course attendance is by invitation only for qualified participants (the
maximum number is 80). Prospective participants should send information in
support of their application (current interests, brief curriculum,
recommendation letters). The application form must be received by May 31,
1993. Notification will be given by June 15, 1993. Participants are
expected to stay for the entire period of ASI.


A registration fee of Italian lire 300.000 is charged only to participants
from the industry (details on payment are sent with final notification).
Participants are expected to stay at Il Ciocco, which provides full board
at a cost of about 95.000 lire per day in double occupancy room (115.000
lire in single room). Full board costs have to be covered directly by the
participants. Participants from NATO countries and East European countries
who do not belong to "for profit" organizations may ask for a contribution
towards boarding expenses. Participants from Greece, Turkey and Portugal
may also ask for a grant for travel expenses. Send to the course director
a motivated letter requesting such a grant (for graduate students, add a
recommendation letter from the thesis advisor).


From: Willard Miller <>
Date: Mon, 1 Mar 93 13:12:31 CST
Subject: Designing a Course in Industrial Mathematics



May 15 - 16, 1993

Supported by a grant from Division of Undergraduate Science,
Engineering and Mathematics Education, National Science Foundation.

With NSF support the IMA and the School of Mathematics at the
University of Minnesota are developing a one year course on
industrial mathematics targeted to juniors and seniors who have
completed vector calculus and are interested in the use of
sophisticated mathematics to solve real-life problems. Initially,
the course is using TeXed lecture notes prepared by Avner Friedman
and Walter Littman which draw on actual industrial problems
presented during the first 5 years of the IMA Seminar on Industrial
Problems and documented in Friedman's books ``Mathematics in
Industrial Problems'' Volumes I-V. (This material will be published
as textbook.) The mathematics foundation is the theory of
ordinary and partial differential equations and this theory is
developed in tandem with the presentation of appropriate industrial
problems which can be modeled via differential equations. Bernardo
Cockburn is writing supplementary notes on the computational aspects
of the industrial problems. The fully developed course can be taught
by a single instructor with a PDE background. (Students would still
need access to a computational laboratory staffed by, say, a TA.)

This meeting is intended for faculty from other colleges and
universities with an interest in offering this or a similar course
in industrial mathematics for undergraduates. It will be devoted
partly to an explanation of the Minnesota experience, partly to
training of potential instructors, and partly to intensive
interaction among the participants to get ideas for improving the
course and developing similar courses.

The registration cost is $16 for the dinner and $10 for the
manuscript. The manuscript for the first 6 chapters of the 7
chapter book will be sent to those who register in advance.
Details about the schedule and a registration form are available
via anonymous ftp at
The registration form should be returned by April 30, 1993.

University of Minnesota
514 Vincent Hall
206 Church Street S.E.
Minneapolis, Minnesota 55455

FAX (612) 626-7370 telephone (612) 624-6066
TeX files of Newsletters, Updates and preprints via
anonymous ftp:
To finger the IMA weekly schedule type:


From: Art Werschulz <>
Date: Tue, 2 Mar 93 10:09:27 EST
Subject: Barcelona Complexity Workshop


There will be a special workshop on Continuous Algorithms and
Complexity in Barcelona, held October 4-6, 1993. A detailed
conference announcement may be obtained via anonymous ftp from The files are in the cacnet subdirectory of the main
anonymous ftp directory, and are called barcelona.tex, barcelona.dvi,

For further information, contact or Do
*not* ask me for further information, since I have none.

Art Werschulz
ATTnet: Columbia University (212) 939-7061
Fordham University (212) 636-6325


From: Paul Messina <>
Date: Sat, 6 Mar 1993 22:14:30 -0800
Subject: Workshop on Programming Tools for Parallel Machines

Villaggio VALTUR, ALIMINI (Otranto, ITALY)
22-25 June, 1993

The purpose of the workshop is to present and discuss the state of the art
and trends of programming tools for parallel machines; it will cover issues
on languages, development and computational models, development environ-
ments. The workshop is aimed at all those interested in parallel processing
and, in general, at whoever needs to acquire a complete and updated picture
of the present and projections of the future.

Speakers: The preliminary program includes talks by:

R. Bisiani, Chairman Programme Committee; Venice University
B. Fadini, Director, Italian Finalized Project
D. Skillicorn, Kingston University, Canada
J. Saltz, ICASE, NASA Langley Research Center, U.S.A.
J. Sipelstein, Carnegie-Mellon University, U.S.A.
J. McGraw, Lawrence Livermore National Laboratory, U.S.A.
D. Walker, University of South Carolina, U.S.A.
B. Bershad, Carnegie-Mellon University, U.S.A.
A. Hey, Southampton University, U.K.
D. May, INMOS, Bristol, U.K.
G. Fox, Syracuse University, U.S.A.
J. Dongarra, Tennessee University & Oak Ridge National Laboratory, U.S.A.
P. Messina, California Institute of Technology, U.S.A.

Villaggio VALTUR, ALIMINI (Otranto, ITALY)
21-22 June, 1993

The Tutorial is addressed to scientific and technical end-users who are
interested in the exploitation of clusters of heterogeneous, independent
computer systems as an alternative to conventional supercomputers or massive-
ly parallel systems. It will cover some widely used tools for supporting
network computing. The speakers include:

Adam Beguelin, Carnegie - Mellon University, U.S.A.
Adam Kolawa, ParaSoft Corporation Pasadena, U.S.A.
Tim Mattson, Yale University & Scientific Computing Associates, Inc., U.S.A.

For further information contact

Via Santa Maria 36 Telex: 500371 - CNUCE

Giovanni Aloisio
Facolta di Ingegneria, Universita di Lecce
Via Provinciale Arnesano - 73100 Lecce (Italy)
Tel.: 0832-620539; 080-242311 - Fax: 0832-625080


Paul Messina
818-584-5917 FAX

(Contact for help with travel, lodging, local arrangements)
Tre Emme Congressi
Via Risorgimento, 4 - 56126 Pisa (Italy)
Tel.: 050-44154/20583 - Fax: 050-500725


End of NA Digest