NA Digest Sunday, June 1, 1997 Volume 97 : Issue 22

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: NA Digest <>
Date: Sun Jun 1 17:42:31 EDT 1997
Subject: NA Digest Calendar

The Netlib Conferences Database is on the Web at:

NA Digest Calendar
Date Topic Place NA Digest #

June 1- 5 Computer Science Education Uppsala, Sweden 38
June 3- 7 ISAAC Congress Newark, DE 47
June 4- 6 Software for Nonlinear Optimization Ischia, Italy 21
June 6- 8 Control, Signals and Image Processing Winnipeg, Canada 21
June 7-11 Large Scale Scientific Computations Varna, Bulgaria 05
June 9 Session on Linear Algebra Winnipeg, Canada 19
June 9-13 Parallel Solution of PDE Minneapolis, MN 19
June 11-13 Intel Supercomputer Users Group Albuquerque, NM 19
June 11-13 IRREGULAR'97 Paderborn, Germany 45
June 16-18 Computer Methods in Water Resources Byblos, Lebanon 35
June 16-18 Mathematical Issues in Geosciences Albuquerque, NM 96:18
June 16-21 Iterative Methods Milovy, Czech Rep. 37
June 18-20 Scientific Computing in EE Darmstadt, Germany 18
June 18-21 Principles + Practice of Parallel Prog. Las Vegas, NV 27
June 22-28 Dirichlet Forms Heraklion, Crete 02
June 23-27 Modeling Workshop Madrid, Spain 22
June 23-28 Course on Nonlinear Hyperbolic Equations Cetraro, Italy 11
June 24-27 Dundee NA Conference Dundee, Scotland 18
June 24-30 Conference Honoring D. K. Faddeev St. Petersburg, Russia 46
June 26-28 Numerical Analysis and Approximation Pretoria, South Africa 44
June 29-.. Nonlinear Dispersive Waves Heraklion, Crete 02
June 30... Lie Groups and Symmetry Analysis Nordfjordeid, Norway 16

July 3- 4 CFD in Minerals, Metal & Power Melbourne, Australia 33
July 3- 8 Curves and Surfaces Lillehammer, Norway 05
July 4- 5 Honor Lothar Collatz Hamburg, Germany 32
July 8-11 Finite Element Method Taejon, Korea 20
July 9-11 Computational Fluid Dynamics Twente, Netherlands 38
July 9-12 Iterative Methods Laramie, WY 36
July 11-12 Randomization and Approximation Bologna, Italy 03
July 13-17 Gordon Conference on HPC and NII Plymouth NH 11
July 13-18 SIAM Annual Meeting Stanford, CA 36
July 13-15 AWM Workshop at SIAM Meeting Stanford, CA 01
July 14-18 Theoretical and Computational Acoustics New York, NY 96:14
July 20-22 Optimization Coimbra, Portugal 20
July 21-23 Symbolic and Algebraic Manipulation Maui, Hawaii 44
July 23-25 Wavelets and Filter Banks Fairfax, VA 22
July 24-26 Applications of Computer Algebra Maui, Hawaii 14
July 24-25 Matrix Methods in Stochastic Models Winnipeg, Canada 47
July 30... ATLAST Linear Algebra Workshop Madison, WI 08

Aug. 4- 8 Large Eddy Simulation Ruston, LA 40
Aug. 10-13 Finite Difference Methods Rousse, Bulgaria 19
Aug. 10-14 Domain Decomposition Boulder, CO 40
Aug. 11-13 Modeling Complex Turbulent Flows Hampton, VA 22
Aug. 11-15 Computational Mathematics Guangzhou, China 48
Aug. 16-17 Matrix Methods for Statistics Istanbul, Turkey 19
Aug. 17-24 Continuum Mechanics Models Zhukovskii, Russia 03
Aug. 20-22 Local to Global Optimization Linkoping, Sweden 05
Aug. 20-22 Automated Timetabling Toronto, Canada 07
Aug. 24-29 IMACS World Congress Berlin, Germany 96:07
Aug. 24-29 Fast Algorithms Berlin, Germany 37
Aug. 26-29 Euro-Par'97 Passau, Germany 43
Aug. 28-31 Stochastic Modelling of Biointeraction Sofia, Bulgaria 47

Sep. 1- 5 Numerical Solution of ODEs Halle, Germany 96:13
Sep. 2 Reduction in Chemical Kinetics Paris, France 22
Sep. 2- 5 Algoritmy'97, Scientific Computing West Tatra, Slovakia 22
Sep. 8-11 Continuum Mechanics Prague, Czech Republic 08
Sep. 8-12 Applied and Computational Mathematics Gramado, Brazil 20
Sep. 8-12 Parallel Computing Technologies Yaroslavl, Russia 06
Sep. 9-12 Boundary Element Method Rome, Italy 41
Sep. 10-12 Computer Arithmetic Lyon, France 41
Sep. 11-13 Algorithm Engineering Venice, Italy 42
Sep. 15-18 Boundary Integral Methods Manchester, England 27
Sep. 15-18 Parallel Computing in Fluid Mechanics Pisa. Italy 15
Sep. 15-19 Scientific Computing & Diff. Eqns. Grado, Italy 21
Sep. 16-19 Visualization and Mathematics Berlin-Dahlem, Germany 09
Sep. 24-26 Dutch Numerical Mathematicians Zeist, Netherlands 18
Sep. 29... ENUMATH 97 Conference Heidelberg, Germany 47

Oct. 6- 9 Optimization and Linear Algebra Qingdao, China 19
Oct. 13-17 Computational Methods & Function Theory Nicosia, Cyprus 16
Oct. 15-17 Large-scale Analysis Symposium Williamsburg VA. 12
Oct. 17-18 London Math Society on PDEs London, England 47
Oct. 24-25 Kalamazoo Matrix Symposium Kalamazoo, MI 13
Oct. 29.... SIAM Applied Linear Algebra Snowbird, Utah 13

Nov. 3- 5 Fourth European PVM/MPI Users' Group Cracow, Poland 13
Nov. 3- 6 Geometric Design Nashville, TN 15
Nov. 5- 8 Hewlett-Packard Computing Cracow, Poland 08


Jan. 3- 6 Approximation Theory Nashville, TN 10
Jan. 5- 9 Matrix Theory Haifa, Israel 19
Jan. 23-24 Honor Olof Widlund New York, NY 08
Jan. 25-27 Discrete Algorithms San Francisco, CA 18
Feb. 9-13 Hyperbolic Problems Zuerich, Switz. 41
Mar. 24-27 Neurocomputing Munich, Germany 15
Apr. 20-23 Interval Methods Nanjing, China 21
June 14-18 Applied Parallel Computing Umea, Sweden 07
June 22-26 Spectral and High Order Methods Tel Aviv, Israel 19
July 1- 3 Optimization Perth, Australia 21
Aug. 18-27 VideoMath Festival Berlin, Germany 12
Aug. 18-27 International Congresses of Math. Berlin, Germany --
Aug. 31... IFIP World Computer Congress Vienna and Budapest 46


From: Anne Greenbaum <>
Date: Fri, 30 May 1997 15:02:43 -0400
Subject: Change of Address for Anne Greenbaum

Beginning in Sept., 1997, I will leave the Courant Institute to take
a professorship in the Math Department at the University of Washington.

My new address will be:

University of Washington
Department of Mathematics
Box 354350
Seattle, WA 98195-4350

Anne Greenbaum


From: Dirk Laurie <>
Date: Fri, 30 May 1997 17:07:26 +0200 (SAT)
Subject: Complex Analogue of Householder Reflections: Summary

My original posting was in NA Digest 97 #18, 4 May 1997. The points
raised were:
1. Typical texts (and therefore courses) in linear algebra introduce
complex matrices only when they are unavoidable.
2. Some results usually derived for real matrices require more than
an H instead of a T when generalized to complex matrices.
3. In particular, the formula H = I - 2*u*u' with u=(x-y)/norm(x-y), and
norm(x)=norm(y), does not have the property H*x=y unless x'*y is real.
I suggested the definition H = (I-P) + w*P where P=u*u' and abs(w)=1
as the proper way to define elementary unitary matrices.

Thanks to Ed Anderson, Andreas Frommer, Bob Funderlic, Nick Higham,
Morten Gulliksson, Richard Lehoucq, Earlin Lutz, Beresford Parlett,
Claus Schneider, Garry Tee and David Watkins for responding.

Main points emerging from the discussion

1. Software packages, and books written by people involved in the
development of software, nowadays use a treatment essentially
equivalent to the one I suggest.
2. Textbooks on linear algebra (in contrast to matrix computations)
are usually satisfied with the usual definition for H.
3. Householder's original book is the one exception on the above
statement, although one respondent confessed "I've always found
Householder's book impenetrable." On the other hand, another
respondent stated "It is a wonderful book that connects the linear
algebra to the numerical linear algebra in a way that is truly
mathematically elegant."

How I would teach it now

[The following presentation is inspired by that in Trefethen and Bau.
However, their treatment is used to derive the usual case, when x'*y
is real and y-p = -(x-p). ]

We want to map x onto y, where ||x|| = ||y||, such that x and y both
project onto the same point p in the space complementary to span{x-y}.
When x and y are distinct, let u=(x-y)/norm(x-y).

Standard formulas for the components give

x = p + u*(u'*x); y = p + u*(u'*y)

Using the relations between x, y, and u we find that

u'*y = -conj(u'*x)

(instead of merely u'*y = - u'*x) leading to

y = x - u*(u'*x + conj(u'*x))
= (x - u*u'*x) - (conj(u'*x)/(u'*x))*u*u'*x
= (I - P - w*P)*x

I'm willing to argue that y is in a complex sense a reflection of x along
the line with direction u, but of course the matrix obtained at the
end does not in the same sense reflect other points along that line.

Personal reminiscences of Householder

[This is a small excerpt from a wonderful letter. I hope the author
of these lines is willing to identify himself and put the whole text,
plus more along these lines, on the Digest!]

> When Householder taught nla (pretty much from his book titled "the
> theory of matrices in numerical analysis") all his matrices were
> complex!, but he was not a programmer and he was interested in
> TEACHING CONCEPTS and liked the generality of complex matrices.
> His book of course has your concern for Hx=y under the proviso
> that x and y are of equal length and x^H y = y^H x.
> I do not know of any other textbook than his that starts off with
> all its matrices complex nor of any that treat Householder transformations
> as complex. ... Householder always called Householder matrices
> "elementary Hermitian matrices."

An alternative to the Householder reflection

The direct rotation introduced by Chandler Davis gives the orthogonal
matrix closest to I that maps a unit vector with top entry gamma (>=0)
and the rest called s into e_1. It is given by [gamma, s'; -s, W]
where W=I-nu*s*s' with nu=1/(1+gamma). This matrix differs from
H defined in the usual way in that the first row is multiplied by -1.

Other cases that also require some care

1. eig(A') == conj(eig(A))
2. det(A') == conj(det(A))

Books and papers that treat the general case

1. G H Golub and C F Van Loan: Matrix Computations (3rd Edition)
2. N J Higham: Accuracy and Stability of Numerical Algorithms, Exercise 18.3
3. A S Householder: The Theory of Matrices in Numerical Analysis (Dover)
4. R Lehoucq: The Computation of Elementary Unitary Matrices,
ACM Tranactions on Mathematical Software Volume 22, Number 4, pp. 393-400
(December 1996);
5. LAPACK User's Guide, page 70 (SIAM)
6. Shukuzawa, O. and Suzuki, T.: Real tridiagonalization
of Hermitian matrices by modified Householder tranformation,
Proc. Japan Acad., Ser A 72, 102-103 (1996)
7. David S. Watkins: Fundamentals of Matrix Computations,
Exercises 3.2.30-3.2.34 on pp. 157-158. ISBN 0-471-61414-9, (Wiley, 1991)

Dirk Laurie


From: Peter Olsen <>
Date: Mon, 26 May 1997 23:30:00 -0400
Subject: Request for Help with Syllabus: Shampine, Allen, and Pruess

I'm writing to ask for suggestions in developing a syllabus for a
one-semester numerical computations course using the text
"Fundamentals of Numerical Computing" by Shampine, Allen, and Pruess.

I'm a 49-year-old graduate student/adjunct instructor with lots
of experience in industrial mathematics and on-the-job-training,
but absolutely no experience in preparing for a formal course.
The course has previously been taught by full-time faculty with
enough experience to present an excellent course with little
formal preparation. My goal is to have a detailed outline in
place before the course starts in September. I am having a
particularly difficult time in deciding how rapidly to pace the
course through the different topics (such as interpolation,
linear systems, root-finding, etc).

Any suggestions will be gratefully accepted. Example syllabi
will spark great rejoicing. Credit will be given in the course
notes and on the course web-page.

My students should be thanking you already.

Peter Olsen


From: Dmitry Golovashkin <>
Date: Fri, 30 May 1997 19:37:30 -0400
Subject: On-Line Optimization System

Dear colleagues,

I would like to announce new on-line optimization system:
Based on lp-solve package (author Michel Berkelaar)
it allows to solve MILP on-line!
You are welcome to visit and test it.

Dmitry V. Golovashkin Chemical Engineering
Carnegie Mellon University
fax : +1 412 268-5229 Pittsburgh, PA 15213 - 3890
e-mail :


From: Tom Peacock <>
Date: Fri, 30 May 1997 14:58:11 +0100
Subject: New Book, The State of the Art in Numerical Analysis

(Special price for NA Digest subscribers. Save 50 pounds. See below.)

The State of the Art in Numerical Analysis
Edited by I. S. Duff, Group Leader of Numerical Analysis, Rutherford Appleton
and G. A. Watson, Chair of Numerical Analysis, Head of Department of
Mathematics and Computer Science, University of Dundee

The 1996 conference on the State of the Art in Numerical Analysis was organized
to provide the numerical analysis community, and users of numerical methods,
with a forum where an account of the important recent developments in the
subject could be presented in a coherent and concentrated way in a manner
accessible to the non-specialist in the sub-area.

It was always intended that this volume extend to a much wider audience the
opportunity given to those who attended the meeting. In the spirit of this
objective, the publishers are now offering the book at a MUCH REDUCED PRICE of
GBP45 (instead of GBP95) which, it is hoped, will enable individual members of
the NA community to buy their own personal copy.

The book contains full versions of all the papers presented, with one exception
on Neural Nets. In addition to recent developments in mainstream topics, linear
algebra, ordinary differential equations, approximation and optimization, it
contains papers devoted to two important application areas, numerical
tomography and image processing.

Compared with the corresponding 1986 volume, some topics show a continuous and
natural development, while others show significant departure from conventional
trends. A recurring theme is the solution of large problems and exploitation of
structure. Underlying many of the developments is the fact that increasingly
complicated and sophisticated problems are now amenable to an increasingly
powerful range of numerical techniques. This is greatly helped by the advent of
excellent computer languages, like MATLAB, and state of the art Fortran
software such as LAPACK, so that the fundamental building blocks for much of
the armoury of a numerical analysis are now readily accessible.

N J Higham: Recent Developments in Dense Numerical Linear Algebra;
I S Duff: Sparse Numerical Linear Algebra: Direct Methods and Preconditioning;
G H Golub and H A van der Vorst: Closer to the Solution: Iterative Linear
H A van der Vorst and G H Golub: 150 Years Old and Still Alive: Eigenproblems;
J M Sanz-Serna: Geometric Integration;
A Stuart: Convergence and Stability in the Numerical Approximation of Dynamical
A Iserles: Beyond the Classical Theory of Computational Ordinary Differential
C T H Baker: Numerical Analysis of Volterra Functional and Integral Equations;
K E Atkinson: The Numerical Solution of Boundary Integral Equations;
G A Watson: Aspects of Approximation with Emphasis on the Univariate Case;
M J D Powell: A Review of Methods for Multivariable Interpolation at Scattered
Data Points;
J Nocedal: Large Scale Unconstrained Optimization;
D F Shanno and E M Simantiraki: Interior Point Methods for Linear and Nonlinear
A R Conn, N I M Gould and Ph L Toint: Methods for Nonlinear Constraints in
Optimization Calculations;
F Brezzi, L P Franca, T J R Hughes and A Russo: Stabilization Techniques and
Subgrid Scales Capturing;
C M Elliott: Approximation of Curvature Dependent Interface Motion;
E S?li and P Houston: Finite Element Methods for Hyperbolic Problems: a
Posteriori Error Analysis and Adaptivity;
K W Morton: Approximation of Multidimensional Hyperbolic Partial Differential
F Natterer: Algorithms in Tomography;
F Guichard and J-M Morel: Partial Differential Equations and Image Iterative

0-19-850014-9, 576 pages, 10 halftones, 36 line figures,
To be published at the end of June 1997 Hardback
Regular price GBP95.00
Special price for NA-Digest subscribers only GBP45.00 (+GBP4.50 for overseas

To pre-order your copy of The State of the Art in Numerical Analysis please
send an email to and we will send you a proforma invoice at
the special discount price.

The special price will also be available to delegates at the 17th Biennial
Conference on Numerical Analysis, Dundee, Scotland, 24-27th June. If you are
planning to attend this meeting and would prefer to reserve a copy and pay for
it at the meeting (no postage charge!), then please send us an email so we can
arrange to send your copy to Dundee.

Best wishes
Tom Peacock
Oxford University Press


From: Jerome Kreuser <>
Date: Mon, 26 May 1997 00:24:40 +0000 (GMT)
Subject: Modeling Workshop with Applications in Electric Energy Systems


IIT/RAMLAB Modeling Workshop

Mathematical Programming Modeling
with Applications in Electric Energy Systems
Using the General Algebraic Modeling System

Sponsored by the Instituto de Investigacion Tecnologica IIT of
Universidad Pontificia Comillas de Madrid in cooperation with the
George Washington University Department of Operations Research -
Research and Applications on Modeling Laboratory, RAMLAB.

Date and location

June 23-27, 1997 in Madrid, SPAIN.

This five-day workshop covers material from a two semester
graduate course in applied modeling. The applications are
focused on electric energy systems but also include related
applications in finance, economic planning, and projects. The
workshop covers theory, algorithms, applications, problems, the
modeling process, and uses the General Algebraic Modeling System
GAMS in a modeling laboratory for studying applications.

A special feature of this workshop is the allocation of an entire
day for the discussion and experimentation with complex and real
applications in the electric power sector currently in use in
electric utilities and developed by IIT.

For more information see:


From: Song Wang <>
Date: Thu, 29 May 1997 09:40:10 +0800
Subject: The Fourth International Conference on Optimization

The Fourth International Conference on Optimization:
Techniques and Applications
July 1-3, 1998, Perth, Australia


Contributions on all aspects of optimization are invited. Only previously
unpublished papers will be considered for presentation at the conference.
All submitted papers will be reviewed by the Technical Program Committee
and the papers presented will be published in the proceedings of the
Participants who wish to contribute a paper are requested to submit three
copies of the paper to the conference secretary:

Dr. Y.H. Leung
Curtin University of Technology
GPO Box U1987, Perth 6845

Last date for submission of paper: September 30, 1997.
Notification of acceptance of paper: November 14, 1997.
Last date for early registration: December 12, 1997.

Further information cab be found at the Web site:


From: Manuel Salas <>
Date: Thu, 29 May 1997 08:03:51 -0500
Subject: Turbulence Modeling Symposium

ICASE/LaRC/AFOSR Symposium on Modeling Complex Turbulent Flows
August 11-13, 1997

The Institute for Computer Applications in Science and Engineering
(ICASE), NASA Langley Research Center and the Air Force Office of
Scientific Research (AFOSR) will be co-sponsoring a Symposium on
Modeling Complex Turbulent Flows on August 11-13, 1997 at the Radisson
Hotel in Hampton, Virginia. The final agenda and registration forms may
be found at the Website:

If you have any questions, please contact:
Emily Todd, Conference Manager
Mail Stop 403, NASA Langley Research Center
Hampton, VA 23681-0001
(757) 864-2175


From: Gil Strang <>
Date: Thu, 29 May 1997 08:24:41 -0400 (EDT)
Subject: Workshop Course on Wavelets and Filter Banks

taught by Gilbert Strang (MIT) and Truong Nguyen (Boston University)
Wednesday-Thursday-Friday July 23-24-25, 1997
George Mason University : Fairfax VA near Washington DC

TEXT: Participants will receive the new textbook (published in 1996)

Wellesley-Cambridge Press, Box 812060, Wellesley MA 02181

This text is already in class use in many EE and mathematics departments.
It was chosen to accompany MATLAB's Wavelet Toolbox, which will be the
simulation software at the Wavelet Workshop. We also have a new

IMAGE CODER by Truong Nguyen (1997)

This will be used at the Workshop and it will be provided only to participants.
We will aim for the right balance of theory and applications. The text
gives an overall perspective of the field - which has grown with amazing
speed. The topics will include

1. Analysis of Filter Banks and Wavelets
2. Design Methods
3. Applications (from Lecturers and Participants)
4. Hands-on Experience with Software (including image coding)

These four key areas will be developed in detail:

1. Analysis

Multirate Signal Processing: Filtering, Decimation, Polyphase
Perfect Reconstruction and Aliasing Removal
Matrix Analysis: Toeplitz Matrices and Fast Algorithms
Wavelet Transform: Pyramid and Cascade Algorithms
Daubechies Wavelets, Orthogonal and Biorthogonal Wavelets
Smoothness, Approximation, Boundary Filters and Wavelets
Time-Frequency and Time-Scale Analysis

2. Design Methods

Spectral Factorization
Cosine-Modulated Filter Banks
Eigenfilters and Quadratic Constrained Least Squares
Lattice Structure
Ladder Structure (Lifting)

3. Applications

Audio and Image Compression, Quantization Effects
Transient Detection and Non-Destructive Evaluation
Digital Communication and Multicarrier Modulation
Text-Image Compression: Lossy and Lossless
Medical Imaging and Scientific Visualization
Image Compression / Image Segmentation / Image Enhancement
Video Compression

4. Simulation Software

MATLAB Wavelet Toolbox
ECG Compression

The goal of the Workshop is to be as useful as possible to all
participants. Please request information by an email message
with subject Workshop to the organizer

Gilbert Strang:

We will reply about the program and tuition cost and housing.
The tuition includes the textbook and software. It will be the same as
in 1995 and 1996 (San Jose, Tampa, and San Diego Workshops). It is
reduced by 50% for graduate students. We are very glad to
answer all questions by email. Our Web sites are

Gilbert Strang Room 2-240 MIT Cambridge MA 02139
617 253 4383 fax 617 253 4358


From: Karol Mikula <>
Date: Sun, 1 Jun 1997 11:18:05 +0200
Subject: Algoritmy'97, Conference on Scientific Computing

ALGORITMY'97 - Conference on Scientific Computing
West Tatra Mountains, Slovakia, September 2 - 5, 1997
Second Announcement

Scientific programme committee:
P. Brunovsky (Bratislava), T.K. Dijkstra (Groningen), W. Jaeger (Heidelberg),
J. Kacur (Bratislava), J. Komornik (Bratislava), K. Mikula (Bratislava)

Topics: modeling of flow in porous medium, simulations of free boundary
phenomena, computational fluid dynamics, modeling of reaction-diffusion
systems, financial and economical modeling, computational geometry, image
processing, scientific visualization.

Invited lectures:
P. Bastian (Stuttgart),
Efficient solution of multiphase flow problems in porous media
E. Baensch (Freiburg),
Adaptive finite element methods - concepts and applications
P. Frolkovic (Erlangen/Bratislava),
Upwinding techniques for convection dominated transport in porous media
D. Hilhorst (Paris), Finite volumes and nonlinear diffusion equations
R.H.W.Hoppe (Augsburg), Adaptive multilevel techniques for solving PDEs
R. Kornhuber (Stuttgart),
Monotone multigrid methods for solving free boundary problems
P. Knabner (Erlangen),
Adaptive finite volume discretization of density driven flows in porous media
S. Kroemker (Heidelberg), Modeling of reaction - diffusion systems
G. H. Meyer (Atlanta), Pricing american options
K. Mikula (Bratislava), Nonlinear diffusion in computer vision
M. Rumpf (Bonn), Visualization of large scale scientific data
M. Paolini (Udine), Numerical methods for geometric evolution of interfaces
A. Schmidt (Freiburg), Simulations of 3D crystal growth
J. Sethian (California-Berkeley), Level set methods
M. Slodicka (Munich/Bratislava),
Finite elements in modeling of flow in porous media; How to describe wells
M. Wierse (Stuttgart), Numerical solution of 3D Navier-Stokes equations

Deadline for registration - June 30, 1997.
Further information and electronic registration form is available at


From: Bracy Elton <>
Date: Thu, 29 May 1997 14:27:22 -0700 (PDT)
Subject: Position at Tera Computer Company

Application and Benchmark Programmer

Program the world's fastest and most widely applicable parallel
computer, proving its performance on existing applications and
exploring new scientific and commercial problem areas previously
overlooked for lack of computational power. Tera's MTA has a flat,
cacheless shared-memory making data-locality a non-issue, and
performance programming merely a matter of exposing enough

Candidates must possess exceptional mathematical and computer problem
solving abilities, and be willing to tackle diverse problems. They
must also function well in a cooperative work environment: individual
initiative and good communication skills are required.

In addition, the following qualifications are desirable:

C and Fortran proficiency
Performance programming experience
Scientific programming experience
Commercial programming experience
Applied mathematics background
Computer science background

Corporate Fact Sheet

Synopsis: Tera Computer Company (Nasdaq: TERA) is creating the first
high performance, general purpose parallel computer
that is both easy to program and scalable. Tera's
Multithreaded Architecture (MTA) will deliver 3-10
times the performance of current similarly priced
supercomputers with peak speeds nearing 256 billion
floating point operations per second (Gflop/s).
Tera systems will be priced from $5 million up, in
configurations of between 16 and 256 processors.

Headquarters: Seattle, Washington <URL>
Approximately 65 employees

Market Tera targets the $1 billion plus worldwide market for
Strategy: very-high-performance computers. Tera plans to
sell its supercomputers to U.S. government agencies,
with follow-on sales to scientific and engineering
applications in industry. According to the market
research firm IDC, the supercomputer market is
expected to grow to $2.4 billion by 1998.

Applications: Complex simulations, advanced computer-aided design,
seismic analysis, computational chemistry, national
security and weather forecasting

Management: Burton J. Smith, Chairman and Chief Scientist
James E. Rottsolk, President and CEO
Brian Koblenz, Vice President of Software
Jerry Loe, Vice President of Hardware
Katherine Rowe, Vice President of Manufacturing

March 1997

If you are interested in applying for a position with Tera, please send
your resume to:
Tera Computer Company
2815 Eastlake Avenue East
Seattle, WA 98102-3027
FAX: 206/325-2433


From: Secretary Support <>
Date: Tue, 27 May 1997 15:37:28 +-200
Subject: Contents, Math. of Control, Signals, and System

Contributed by Jan H. van Schuppen


Volume 9, Number 4

P. Dai Pra, L. Meneghini and W.J. Runggaldier,
Connections between stochastic control and dynamic games,
MCSS 9 (1996), 303-326.

T.I. Seidman and Jiongmin Yong,
How violent are fast controls? -II,
MCSS 9 (1996), 327-340.

V.E. Benes and R.J. Elliott,
Finite-dimensional solutions of a modified Zakai-equation,
MCSS 9 (1996), 341-351.

P. Fitzpatrick,
On the scalar rational interpolation problem,
MCSS 9 (1996), 352-369.

Jie Chen and S.S.-T. Yau,
Finite-dimensional filters with nonlinear drift VI:
Linear structure on $\Omega$,
MCSS 9 (1996), 370-385.

O. Toker and H. Ozbay,
Complexity issues in robust stability of linear
delay-differential systems,
MCSS 9 (1996), 386-400.

Information on MCSS including tables of contents is
available at its home pages:
- -
- -

Address for submissions:
J.H. van Schuppen (Co-Editor MCSS)
P.O.Box 94079
1090 Gb Amsterdam
The Netherlands


From: Edward Sisson <>
Date: Wed, 28 May 97 08:39:06 EST
Subject: Contents, SIAM Journal on Matrix Analysis and Applications

SIAM Journal on Matrix Analyis and Applications
Volume 18, Number 3, JULY 1997

The Minimum Eigenvalue of a Symmetric Positive-Definite Toeplitz Matrix and
Rational Hermitian Interpolation
Wolfgang Mackens and Heinrich Voss

Estimating the Attainable Accuracy of Recursively Computed Residual Methods
Anne Greenbaum

Fast Nested Dissection for Finite Element Meshes
Shang-Hua Teng

An Efficient Implementation of the Nonsymmetric Lanczos Algorithm
David Day

On Computing Stable Lagrangian Subspaces of Hamiltonian Matrices and Symplectic
Wen-Wei Lin and Chern-Shuh Wang

The Matrix Sign Function Method and the Computation of Invariant Subspaces
Ralph Byers, Chunyang He, and Volker Mehrmann

Implicitly Restarted Krylov Subspace Methods for Stable Partial Realizations
Imad M. Jaimoukha and Ebrahim M. Kasenally

A Geometric Approach to Perturbation Theory of Matrices and Matrix Pencils. Part
I: Versal Deformations
Alan Edelman, Erik Elmroth, and Bo Kagstrom

On the Shape of the Symmetric, Persymmetric, and Skew-Symmetric Solution Set
Gotz Alefeld, Vladik Kreinovich, and Gunter Mayer

An Analysis of Spectral Envelope Reduction via Quadratic Assignment Problems
Alan George and Alex Pothen

Perturbation of Eigenvalues of Preconditioned Navier-Stokes Operators
Howard C. Elman

Extension of Isometries in Finite-Dimensional Indefinite Scalar Product Spaces
and Polar Decompositions
Yuri Bolshakov, Cornelis V. M. van der Mee, Andre C. M. Ran, Boris Reichstein,
and Leiba Rodman

Perturbation Analyses for the QR Factorization
Xiao-Wen Chang, Christopher C. Paige, and G. W. Stewart


End of NA Digest