NA News Digest Sunday, December 16, 1990 Volume 90 : Issue 44 Today's Editor: Cleve Moler

Today's Topics:


From: Cleve Moler <>
Date: Sun Dec 16 10:12:00 PST 1990
Subject: *** Please send submissions to Oak Ridge, not Stanford ***

NA News Digest Now Coming from Oak Ridge

If you look at the mail header for this week's NA News Digest,
you should see that it is coming from Oak Ridge National Laboratory,
and not from Stanford University. As we announced in last week's
digest, all NA-NET services are being moved to Oak Ridge.

Thanks, again, to Gene Golub and Mark Kent who established NA-NET
at Stanford, to Jack Dongarra and Bill Rosener who are now
handling things at Oak Ridge and the University of Tennessee,
and to Sequent Computers, who have provided the na-net and netlib
machine at Oak Ridge. (The Internet mail addresses ""
and "" are aliases for "".)

As the comments in the mail header now indicate, submissions for the
NA New Digest should be mailed to

Information on NA-NET, including how to join and how to send
individual mail to other members, can be obtained by sending mail to

It doesn't matter what the mail message to actually says; you
will automatically get a reply consisting of last week's annoucement.

-- Cleve Moler, or


From: C. R. Katholi <>
Date: Tue, 11 Dec 90 10:38:00 CST
Subject: Numerical Methods for Delay Differential Equations

I am interested in any pointers to numerical methods for the solution
of delay differential equations. Any information about available computer
codes would also be appreciated.

Contact: Charles Katholi
Department of Biostatistics and Biomathematics
University of Alabama at Birmingham
UAB Station
Birmingham, Al 35294

Thank you in advance for ANY help.


Date: Mon, 10 Dec 90 11:28:49 EDT
Subject: A Conference on Multigrid Methods

The George Washington University (Washington, DC) will host
an NSF/CBMS sponsored regional research conference on Multigrid
and Multilevel Adaptive Methods for Partial Differential Equations
from June 24 to June 28, 1991.

The principal lecturer will be Stephen F. McCormick of the
University of Colorado at Denver who will deliver a sequence
10 one-hour lectures. Demonstrations and workshop sessions will
also be organized. Additional lectures may be given by the
invited participants.

Participation is open to research scientists and graduate students
involved in activities that would profit from the lectures and
interaction the conference would provide. Participation of graduate
students and minorities is strongly encouraged.

A limited amount of support for travel and subsistence is
available. Applications for participation and possible support
should be made by March 1, 1991. To participate, with
or without support, please send a short vita and a brief
statement of interest.

For further information contact the Conference Director:

Professor Murli M. Gupta
Department of Mathematics
The George Washington University
Washington, D.C. 20052
Phone: 202-994-4857
FAX: 202-994-0458
E-mail: MMG@GWUVM.bitnet


From: Jeffrey Scroggs <>
Date: Tue, 11 Dec 90 15:20:28 EST
Subject: Fifth Conference on Domain Decomposition Methods

Fifth Conference on Domain Decomposition Methods
sponsored by SIAM

Deadline: December 18, 1990

Dear Colleagues,

This is the final call for abstracts for the Fifth Conference on
Domain Decomposition that will be held May 6-8, 1991 in Norfolk, VA.

This conference will bring together specialists in areas such as
PDEs, numerical analysis, computer science, and parallel computing
as well as engineers from the aerospace, oil and chemical industries.
Anyone involved the numerical solution of large systems of partial
differential equations on state-of-the-art computer systems should
find the meeting of interest.

Invited Speakers

Maksymilian Dryja, Warsaw University
Roland Glowinski, University of Houston
Gene Golub, Stanford University
William Gropp, Argonne National Laboratory
Yuri Kuznetsov, USSR Academy of Sciences
Joseph Pasciak, Brookhaven National Laboratory
Anthony Patera, Massachusetts Institute of Technology
Jacques Periaux, Avions Marcel Dassault and INRIA
Alfio Quarteroni, University of Minnesota
Olof Widlund, Courant Institute of Mathematical Sciences

We encourage contributions in the following areas:

* Numerical Analysis of Domain Decomposition Methods
* Hyperbolic Problems and Computational Fluid Dynamics
* Coupling of Domain Decomposition and Multilevel Methods
* Interface Conditions for Domains that have Different Governing Equations
* Multidomain Spectral Methods
* Engineering Applications
* Parallel Implementations

There will be the following three types of presentations: 25-minute lectures,
15-minute lectures, and poster presentations. To give a 15-minute lecture or
a poster, submit a brief abstract (approx. 100 words). To be considered for a
25-minute lecture, submit an extended abstract (approx. 1000 words) in
addition to the brief abstract. Persons submitting extended abstracts will
be notified of their classification by January 21, 1991.

Abstracts should be sent to:

3600 University City Science Center
Philadelphia, PA 19104-9800


Deadline for abstracts is December 18, 1990. If you need an extension
of the deadline, please contact me at or (804)864-2174.

Robert G. Voigt
Conference Chairman


From: Volker Mehrmann <UMATF108%DBIUNI11.BITNET@Forsythe.Stanford.EDU>
Date: Fri, 14 Dec 90 10:44:13 CET
Subject: Preprints Available from Mehrmann

The following 1990 preprints are available from
Volker Mehrmann, Fak. f. Mathematik, Univ. Bielefeld,
Postfach 8640, D-4800 Bielefeld 1, FRG

G. Ammar , V. Mehrmann, On Hamiltonian and Symplectic Hessenberg Forms ,
Sonderforschungsbereich 343 `Diskrete Strukturen in der Mathematik',
Universitaet Bielefeld, Report 90-002, to appear in Linear Algebra
and its Applications

A. Bunse-Gerstner, R. Byers, V. Mehrmann, Numerical Methods for
Simultaneous Diagonalization, Sonderforschungsbereich 343,
`Diskrete Strukturen in der Mathematik', Universitaet Bielefeld,
Report 90-009, submitted.

A. Bunse-Gerstner, L. Elsner, The Schur parameter form for the solution of the
unitary eigenvalue problem, Sonderforschungsbereich 343,
`Diskrete Strukturen in der Mathematik', Universitaet Bielefeld,
Report 90-018, to appear in Linear Algebra and its Applications

L. Elsner, V. Mehrmann, Convergence of Block--Iterative Methods for
Linear Systems Arising in the Numerical Solution of Euler Equations,
Sonderforschungsbereich 343, `Diskrete Strukturen in der Mathematik',
Universitaet Bielefeld, Report 90-036, submitted for publication

T. Geerts, Infinite-horizon linear quadratic control: A state of the art,
Sonderforschungsbereich 343, `Diskrete Strukturen in der Mathematik',
Universitaet Bielefeld, Report 90-038

A. Bunse-Gerstner, R. Byers, V. Mehrmann, N.K. Nichols,
Numerical Computation of an Analytic Singular Value Decomposition of a
Matrix Valued Function, FSP Mathematisierung, Universitaet Bielefeld,
Materialien LIX, 1990, submitted for publication

P. Kunkel, V. Mehrmann, Smooth factorizations of matrix valued
functions and their derivatives, Preprint 63, Institut f. Geometrie und
Praktische Mathematik, RWTH Aachen, Templergraben 55, 5100 Aachen, FRG,
submitted for publication

T. Geerts, V. Mehrmann, Linear differential equations with constant
coefficients: A distributional approach, Sonderforschungsbereich 343,
`Diskrete Strukturen in der Mathematik', Universitaet Bielefeld,
Report 90-0??, submitted for publication

W.J. Beyn, Numerical methods for dynamical systems, to appear in Proceedings
of the SERC Summer-School at Lancaster (UK), Oxford University Press 1990


From: Iain Duff <>
Date: Fri, 14 Dec 90 14:35:26 GMT
Subject: CERFACS Short Course on Parallel Programming

(European Centre for Research and Advanced Training in
Scientific Computation)

Will Organize

An Introduction to Parallel Computing
A three-day advanced short course
Monday 22nd April - Wednesday 24th April 1990
Toulouse, France


This course on parallel computing will be held at CERFACS, Toulouse from
April 22nd to 24th, 1991 under the auspices of CERFACS.

The course is aimed at scientists in industry and academia who wish to
learn about parallel computing and familiarize themselves through hands-on
experience with a range of parallel computers.

On the first day, some introductory lectures on vector and parallel
architectures will be followed by a more detailed description of the
parallel computers available to students on the course. These will include
a BBN TC2000, an Alliant FX/80 and FX/2812, a TMC Connection Machine CM-2,
an INTEL hypercube, a CRAY 2 and an IBM 3090/VF.

The second day will examine techniques for programming these machines and
tools available to assist in their use. These lectures will be followed by
a discussion of basic numerical algorithms and software.

The final day will address more large scale computing on such
architectures, including the solution of sparse equations, and will provide an
opportunity for students to run their own codes on the machines.

Throughout the course, there will be a strong emphasis on hands-on
experience and it is the intention that students will finish the course
with a basic knowledge of parallelism and how to appreciate, compare, and
use a range of different parallel architectures.


Participants are advised to register before April 1 by Email, fax or phone to

Dominique BENNETT,
CERFACS, 42 Ave Gustave Coriolis,
Tel : (33) 61 07 96 96,
Fax : (33) 61 07 96 13
EMail :


Dr P.Amestoy is a post-doc in the Parallel Algorithms Group at CERFACS
whose primary interest is the solution of sparse linear equations using
multifrontal methods. He is also interested in the application of these
techniques in CFD calculations.

Dr M. Arioli is a senior scientist in the Parallel Algorithm Group and a
Research Scientist of CNR. His principal interests are in error analysis,
parallel computing, and sparse matrices.

Y-H De Roeck is a graduate student at CERFACS working in the solution of
problems in linear and nonlinear elasticity using domain-decomposition.

Professor J.Dongarra has a joint appointment as a Distinguished Scientist
at the University of Tennessee and Oak Ridge National Laboratory. His
principal interests are in linear algebra and supercomputing and he is
playing a leading role in the LAPACK project.

Professor I.Duff is the Project Leader of the Parallel Algorithm Group at
CERFACS and is Group Leader of Numerical Analysis at the Rutherford
Appleton Laboratory and a visiting Professor of Mathematics at the
University of Strathclyde. He is the principal editor of the IMA Journal
of Numerical Analysis and his main interests are in sparse matrices and

Professor W.Jalby has a joint position with INRIA and the University of
Rennes. He is a principal editor of the International Journal of High
Speed Computing and specializes in parallel programming and computation

D.Levine is a graduate student and member of the computer support team at
the ACRF at Argonne National Laboratory. He is the main organizer of
their courses in parallel computing.

D. Ruiz is a graduate student in the parallel Algorithm Group at CERFACS
working in the solution of sparse systems by block iterative methods.

Dr L.Valdettaro is a post-doc in the Turbulence Group at CERFACS working on
the study of three-dimensional models of turbulence and their
implementation on parallel architectures


From: Iain Duff <>
Date: Fri, 14 Dec 90 14:36:10 GMT
Subject: Positions at CERFACS




CERFACS aims at developing algorithms and new tools for parallel
computation with a view to applying them to large scale scientific
problems. Another major mission involves advanced training in scientific
computation. Based in Toulouse, France, CERFACS has researchers and
engineers from many countries in Europe and elsewhere.

CERFACS offers, for October 1991,
Post-Doctorate and PhD grants.


* Parallel Algorithms and numerical computation
* Direct Simulation on Instabilities and Turbulence
* Computational Aerodynamics and numerical physical modelling of
* Visualization Methods especially for computational fluid mechanics
* Global Climate Modelling and Numerical Simulation of Ocean-Atmosphere

For more information, contact:
42 Avenue Gustave Coriolis, 31057 TOULOUSE Cedex (France)
Tel.: (33) 61 07 96 96 - Fax: (33) 61 07 96 13


From: Robert Feldman <>
Date: Mon, 10 Dec 90 09:03:50 EST
Subject: Second Symposium on High Performance Computing

October 7-9, 1991
Montpellier, France

Organized by
CNUSC: Centre National Universitaire Sud de Calcul
CNSF: Cornell National Supercomputer Facility
INRIA: Institut National de Recherche en Informatique et en Automatique

This is the second in a series of biennial symposia, the goal being an
overview of innovations in the high performance computing field. During
the last decade, intensive computing (parallelism, vectorization, and
scalar) has drawn a large audience in the world of engineers and
researchers, particularly in the field of classical computations. This
has opened several "new" perspectives which were impossible to realize
or conceptualize previously.

The main purpose of this meeting is to bring scientists from various
fields together in order to investigate areas such as Architectures of
Supercomputers, Compilers, Algorithms, Computational Methods, Numerical
Applications, etc.

You are invited to submit a paper on one of the following topics:

1) Architectures and parallel algorithms.
2) Tools and numerical methods oriented to parallel/vector computations.
3) Applications on supercomputers:

E-mail: SHPC@FRMOP11.Bitnet

Scientific Organizing Committee

F. Daumas, CNUSC (France)
M. Durand, IBM (France)
F. El Dabaghi, INRIA (France)
R. Feldman, CNSF (USA)

Scientific Program Committee

Z. Barzilai, IBM Yorktown Heights (USA)
C. Basdevant, Universite Paris XIII / ENS Paris (France)
C. Benoit, Universite Montpellier II (France)
L. Brown, Cornell University (USA)
J-L. Delhaye, CNUSC (France)
J. Demaille, Universite Montpellier I (France)
I. Duff, RAL (UK) / CERFACS (France)
J. Erhel, INRIA-IRISA (France)
L. Fezoui, INRIA (France)
P.J. van der Houwen, CWI (Netherlands)
W. Jalby, INRIA (France)
M. Kalos, Cornell Theory Center (USA)
E. Krause, RWTH Aachen (Germany)
D. Laforenza, CNUCE (Italy)
P. Leca, ONERA (France)
O. Mc Bryan, U. of Colorado, Boulder (USA)
F. Marcotorchino, CEMAP IBM ( France)
O. Pironneau, U. Paris VI / INRIA (France)
R. Scarborough, Palo Alto Sci Ctr IBM (USA)
P. Sguazzero, ECSEC IBM Europe (Italy)
K. Stuben, GMD (Germany)

For more information:

E-mail: SHPC@FRMOP11.Bitnet

Robert L. Feldman, Coordinator
International Supercomputing Alliance
Cornell National Supercomputer Facility
Theory Center Building
Ithaca, New York 14853-3801
(607) 254-8686


From: Hatem el-Sebaaly <!caen!!>
Date: 11 Dec 90 22:38:05 GMT
Subject: World Congress of Nonlinear Analysts

Melbourne, Florida USA
August 19-26, 1992

FIRST ANNOUNCEMENT: The Global Scientific Advisory Committee (GSAC) is
pleased to announce that the First World Congress of Nonlinear Analysts will
be held from Wednesday, August 19 through Wednesday, August 26, 1992 at Florida
Institute of Technology in Melbourne, Florida, USA.

The interest in mathematical analysis of nonlinear phenomena has led to
increased interdependency and cooperation among scientists across traditional
disciplinary barriers, and this collaboration has proved to be rewarding
and stimulating. Also, industrial recognition of using nonlinear mathematical
models for solving problems of technology is increasing because of the fast development of large-scale computers. It is, therefore, clear that the attempt to
understand the nonlinear world will dominate large parts of mathematics in
most areas of science and technology in future years. Thus, there is a
definite need to have a forum, on a global basis, to forge unity in diversity
and for bringing more cooperation and collaboration into the world community
of nonlinear analysts. It is with this spirit we feel that it is desirable
to have, periodically, a world congress.

R. Conti, K.V. Frolov, J. Hale, K. Ito, W. Jager, T. Kato, M.A. Krasnoselskii,
J. Kurzweil, O. Ladyzenskaya, V. Lakshmikantham, J.L. Lions, Li Ta-Tsien (Li
Da-Qian), G.I. Marchuk, V.M. Matrosov, Yu. A. Mitropolsky, L. Nirenberg,
C. Olech, I. Prigogine, P.A. Samuelson, A.A. Samarsky and M.P. Singh.

SCIENTIFIC PROGRAM: There will be approximately 20 invited expository
addresses covering recent developments in major areas of Nonlinear Analysis
and about 120 invited lectures divided into the following six main sections
that will have subsections:

1. Nonlinear Operators
2. Calculus of Variations and Optimization; Control Theory; and
Systems Analysis
3. Numerical and Computational Aspects; Mathematical Modelling
4. Dynamic Systems
5. Stochastic Analysis and Applications
6. Real-World Applications

Professor V. Lakshmikantham
Chairman, GSAC
Florida Institute of Technology
Department of Applied Mathematics
Melbourne, Florida 32901 USA

Telephone: (407) 768-8000 ext. 8091 or 7412
FAX: (407) 984-8461


From: Jon Kettenring <>
Date: Wed, 12 Dec 90 17:49:28 -0500
Subject: Interface'91 Call For Papers & Attendees


23rd Symposium on the Interface:
Computing Science and Statistics

Critical Applications of Scientific Computing:
Biology, Engineering, Medicine, Speech...

APRIL 21-24, 1991 - Seattle Sheraton Hotel, SEATTLE, WASHINGTON
Sponsored by the Interface Foundation of North America

Hosted by Bellcore


"Opportunities for Statisticians and
Computer Scientists in Biology"


Relational Databases - Joe Hill
Parallel Computing - Bill Eddy and Mark Schervish
Realistic Rendering - Tony DeRose


Dynamic Statistical Graphics
Computationally Intensive Methods for Discrete Data
Multivariate Statistics and Visualization for Labelled Point Data
Massive Data Bases
Computational Problems in Biomedical Imaging
Uncertainty and Graphical Models
Computational Genetics
Computational Problems in Satellite Image Analysis
Computational Methods in Spatial Statistics
Speech and Language
Computer Modeling, Experimental Design and Data Analysis
Scientific Computing Problems in the Aircraft Industry
Computing Problems in Environmental and Industrial Statistics
Data Visualization and Sonification
Neural Networks
Engineering Applications of Computing-Intensive Methods
Statistical Computing Environments for the 21st Century


Sessions will include invited and contributed papers,
tutorials, and a special session of invited posters, videos,
and demos. Authors who wish to give a contributed paper or
to be considered for the poster/video/demo session should
submit an abstract by January 15, 1991. Abstracts should
not exceed one-half page (text 6.5 inches wide by 4 inches)
including lines for title, author(s), and address.
They may be sent as camera-ready paper copy (2 copies,
please) or email (PostScript, troff, or TeX/LaTeX). All
abstracts will be included in the program brochure.


Mary Ellen Bock, Andreas Buja, William DuMouchel, Nick
Fisher, Gene Golub, Joe Hill, Elaine Keramidas, John
McDonald, John Nash, Daryl Pregibon, Werner Stuetzle,
Michael Tarter, Luke Tierney, Paul Tukey, Paul Young


Diane Duffy, Selma Kaufman, Elaine Keramidas, Jack Lee,
Allen McIntosh, Elaine Sieff, Linda Uhler

Abstracts and inquiries should be sent to:

mail: Interface '91
Jon Kettenring, Program Chair
Bellcore, 2Q-326
445 South Street
Morristown, NJ 07962-1910
phone: (201)829-4398


End of NA News Digest

From num Sun Dec 16 22:13 EST 1990
Original-address: From surfer.EPM.ORNL.GOV!moler Sun Dec 16 20:55:30 0500 1990 remote from pyxis
Received: by pyxis; Sun Dec 16 22:13 EST 1990
Received: by; Sun Dec 16 22:13 EST 1990
Received: by surfer.EPM.ORNL.GOV (5.61/1.34)
id AA08427; Sun, 16 Dec 90 20:55:30 -0500
Date: Sun, 16 Dec 90 20:55:30 -0500
From: moler@surfer.EPM.ORNL.GOV (Cleve Moler)
Message-Id: <9012170155.AA08427@surfer.EPM.ORNL.GOV>
To: na-comb@surfer.EPM.ORNL.GOV

Subject: Correction to NA News Digest, vol. 90, Issue 44
From: The NA-NET <>

OOPS! I made a couple of dumb mistakes in the NA New Digest sent out
earlier today. Please make the following changes in the first article.

As the comments in the mail header now indicate, submissions for the
NA New Digest should be mailed to
should be:

Information on NA-NET, including how to join and how to send
individual mail to other members, can be obtained by sending mail to
should be:

Sorry 'bout that.

-- Cleve