### Today's Editor:

- Cleve Moler
- The MathWorks, Inc.
- moler@mathworks.com

- References for SIMD Linear Algebra
- Seeking a Better Expansion
- Looking for the Second Edition of Lawson and Hanson
- A Great Bargain
- C++ for Numerical Analysis
- Hellenic Mathematical Society, 75 Years
- 1993 PVM Users' Group Meeting
- Numerical Methods for the Solution of the Maxwell Equations
- ACM-SIAM Symposium on Discrete Algorithms
- Position at CERCA, Montreal
- Contents, SIAM Math Analysis
- Contents, Acta Numerica

-------------------------------------------------------

From: Glenn R. Luecke <GM.GRL@ISUMVS.IASTATE.EDU>

Date: Mon, 12 Apr 93 08:44:29 CDT

**Subject: References for SIMD Linear Algebra**

I am looking for references for papers dealing with parallel linear algebra

routines (especially for nonsparse LU factorization, Cholesky factorization,

SVD and eigensystem solvers) for SIMD computers, such as Thinking Machines'

CM-2 and MasPar's MP-1/MP-2 computers. So far I have not found any such

references and would greatly appreciate it if anyone could provide me with

such. Thanks!

Glenn Luecke

GM.GRL@ISUMVS.IASTATE.EDU

291 Durham Center, Iowa State University, Ames, IA 50011

------------------------------

From: Stewart Levin <salevin@dal.mobil.com>

Date: Mon, 12 Apr 93 13:08:52 CDT

**Subject: Seeking a Better Expansion**

A problem arising in exploration geophysics is that of dereverberation.

We probe an earth structure X with an impulsive signal and receive a

response U=X(1+X)^-1 due to (free surface) reverberation. One way to

recover X from U is to form

U(1-U)^-1 = U + U^2 + U^3 + U^4 + ...

which in the limit recovers X. A problem that arises in practice is

that (1-U)^-1 is not practical to compute directly and so we employ

the MacLaurin expansion instead. Taking partial sums of the

MacLaurin series has a bad side effect of boosting later reverbera-

tions as it suppresses earlier ones, e.g. partial sums look like:

S1 = X - X^2 + X^3 - X^4 + X^5 - X^6 + X^7 - X^8 + ...

S2 = X - X^3 + 2X^4 - 3X^5 + 4X^6 - 5X^7 + 6X^8 + ...

S3 = X - X^4 + 3X^5 - 6X^6 +10X^7 -15X^8 + ...

S4 = X - X^5 + 4X^6 -10X^7 +20X^8 + ...

S5 = X - X^6 + 5X^7 -15X^8 + ...

S6 = X - X^7 + 6X^8 + ...

S7 = X - X^8 + ...

S8 = X + ...

Can anyone come up with an alternative expansion, e.g. in terms of

polynomial families or product representations, which has the desirable

property that the coefficients of the Taylor expansion of the partial

sums (or products) is always less than 1 in magnitude?

-- stew@hanauma.stanford.edu

-- salevin@dal.mobil.com

------------------------------

From: Chuck Lawson <clawson@math.Jpl.Nasa.Gov>

Date: Tue, 13 Apr 93 9:22:29 PDT

**Subject: Looking for the Second Edition of Lawson and Hanson**

To: readers of the na.digest --

I would like to buy a copy of the Second Edition of

"Solving Least Squares Problems" by Lawson and Hanson,

published by Prentice-Hall. It is now out of print.

Please contact me if you have a copy you are willing to

sell.

-- Charles Lawson, clawson@math.jpl.nasa.gov

------------------------------

From: Gene Golub <golub@sccm.Stanford.EDU>

Date: Tue, 13 Apr 93 21:58:27 PDT

**Subject: A Great Bargain**

Dover Publications has just re-issued Herb Keller's "Numerical Methods for

Two-Point Boundary Value Problems" The cost is $US 9.95. (That's not a

typographical error!) The volume also includes Keller's monograph published

by SIAM and several of Keller's paper. Coming soon: Isaacson and Keller.

Thanks, Dover. You are doing a terrific job for our community.

Gene Golub

------------------------------

From: Walter Figueiredo Mascarenhas <walter@ime.unicamp.br>

Date: Fri, 16 Apr 93 16:20:03 BSC

**Subject: C++ for Numerical Analysis**

I would like to hear the impressions from people using C++ for

Numerical Analysis. Is it true that C++'s (and C's) numerics are not

as good/accurate/stable as FORTRAN's? How good are the libraries

Linpack.h++, from Rogue Wave, and M++, from Dyad? Do you know other

(supported and well documented) libraries?

If you think that FORTRAN is THE language for numerical analysis and

would not waste your time with C++, please let me know why.

Thanks for your help,

Walter Mascarenhas

walter@ime.unicamp.br

------------------------------

From: Hellenic Math Society <DEMETRI@aueb.ariadne-t.gr>

Date: Wed, 14 APR 93 16:44 GMT

**Subject: Hellenic Mathematical Society, 75 Years**

SCIENTIFIC EVENTS OF THE HELLENIC MATHEMATICAL SOCIETY

FOR THE COMPLETION OF 75-YEARS FROM ITS ESTABLISHMENT

The Hellenic Mathematical Society (HMS) on the occassion of the

completion of 75-years of active multi-disciplinary scientific

contribution in the Hellenic and International Mathematical Community, is

organizing a series of scientific events during the year 1993.

These celebratory scientific activities of HMS include:

(I) MATHEMATICAL ONE-DAY MEETINGS (HMS-75-MH)

A series of scientific meetings and discusions with general and

specific subjects of Mathematics and its applications. These scientific

(one-day) meetings will take place in Athens at regular time-periods

during the year 1993. The three distinguished speakers in such a meeting

will be selected from the following mathematical categories:

(i) Members of the Academy of Athens and ex-Presidents of the HMS.

Subject of their talks: the contribution of Greek Mathematicians in the

development of the modern Mathematical Science, the activities of the HMS

etc.

(ii) Heads of the Departments of Mathematics of the Hellenic

Universities and Technical Universities as well as distinguished

Researchers-Professors of Academic Institutions. Subject of their talks:

Modern programs of mathematical studies and their applications, present

research activities of Departments of Mathematics, participation in

Hellenic and International Research Programmes etc.

(iii) Presidents and national representatives of the Mathematical

Societies of the following countries:

United States of America, C.I.S. (Academy of Science, Moscow),

Japan, China, Britain, Germany, France,

as well as national representatives from other European countries. Subject

of their talks: present activities of the Scientific Societies of their

countries, participation in international research programmes, development

of international scientific relationships, personal research interests etc

(II) HMS SCIENTIFIC RESEARCH AWARD (HMS-75-D )

The HMS announces a scientific research award for the most outstanding

Doctoral Thesis in Mathematics and its applications written by Greek

scientist in the period 1990-1993.

This scientific distinction will be awarded to the eminent Greek Ph.D.

author, whose the doctoral programme has been completed in a Hellenic

University (or corresponding Academic Institution abroad) in a special

ceremony of the HMS near the end of 1993, after the final decision of the

HMS-75-D Evaluation Committee.

The prospective candidates for the above HMS Scientific research award

should sent three copies of their Doctoral Thesis (Ph.D), written in Greek

or English language, and a short CV in the following address:

Hellenic Mathematical Society

HMS-75-D Evaluation Committee

34 Panepistimiou Street

GR- 106 79, Athens

G r e e c e

The deadline for the submission of entries is July 31st, 1993.

(III) SPECIAL VOLUME FOR THE 75-YEARS of HMS (EME-75-EE)

In the framework of its foundation purposes and with the occassion of

the completion of 75-years, the HMS will publish a collective scientific

volume with main subject:

"The contribution of the ancient Greek Mathematicians in the

development of the modern Mathematical Science and its applications "

In this work have been invited to participate with their distinguished

contributions members of the National Mathematical Committee (Academy of

Athens), University Professors and eminent Research Workers. Those who

would like to submit suitable work for inclusion in this HMS-75 volume

please write to HMS address given above.

Professor Elias A. Lipitakis

President of Hellenic Mathematical Society

------------------------------

From: Jack Dongarra <dongarra@cs.utk.edu>

Date: Wed, 14 Apr 93 12:56:57 -0400

**Subject: 1993 PVM Users' Group Meeting**

1993 PVM Users Group Meeting

Knoxville, Tennessee

May 10 - 11, Knoxville Hilton, Knoxville, TN

The first PVM Users' Group Meeting will be held May 10th and 11th,

1993, at the Downtown Knoxville Hilton, Knoxville, TN. The meeting

will provide a chance for users of PVM to meet each other and

members of The PVM Team.

The conference will include:

* Invited talks by experts in fields of interest to PVM users.

* Contributed papers from the PVM user community.

* Technical presentations by the developers of PVM.

* Social activities.

Call For Papers/Posters

We welcome abstracts of papers and posters for possible presentation

at the Users' Group meeting.

Abstracts should be approximately 300 words in length.

The deadline for submitting abstracts is April 23, 1993. Abstracts

will be reviewed and authors notified of their acceptance by April 30.

Please submit abstracts to:

Jan Jones

Computer Science Department

University of Tennessee

Knoxville, TN 37996

phone: 615-974-8295, fax: 615-974-8296

or to: pvmug@cs.utk.edu

Additional information on the Users' Group meting can be found in netlib.

Send email to netlib@ornl.gov and in the mail message type:

send pvmug93 from pvm

------------------------------

From: Rainald Lohner <lohner@seas.gwu.edu>

Date: Thu, 15 Apr 93 11:43:25 EDT

**Subject: Numerical Methods for the Solution of the Maxwell Equations**

SECOND INTERNATIONAL CONFERENCE AND WORKSHOP ON APPROXIMATIONS AND

NUMERICAL METHODS FOR THE SOLUTION OF THE MAXWELL EQUATIONS

Key Bridge Marriott

Washington, DC

October 25-27, 1993

CALL FOR PAPERS

The organizing committee is pleased to announce the

SECOND INTERNATIONAL CONFERENCE AND WORKSHOP ON APPROXIMATIONS

AND NUMERICAL METHODS FOR THE SOLUTION OF THE MAXWELL EQUATIONS.

The aim of the Conference is to review progress and assess new techniques

for the numerical solution of the Maxwell equations, with particular

emphasis on the applicability for Radar Cross Sections (RCS) calculations

of perfectly and partially conducting objects. The Conference will feature

invited speakers as well as contributed papers.

The aim of the Workshop is to compare the different numerical methodologies

currently employed for RCS calculations. Comparisons in terms of accuracy

and efficiency will be conducted on a set of 24 selected test cases.

Participation to the Workshop is open to anybody willing to compute and

document at least two of the mandatory test cases which consist of different

geometries.

The Conference and Workshop will be held in Washington, DC, with the help of

CMEE/SEAS-GWU (US) and GAMNI-SMAI (Europe).

Organizing Committee:

R. Glowinski (Univ. of Houston), R. Lohner (The George Washington University,

Co-Chair), J. Periaux (GAMNI-SMAI, Co-Chair), V. Shankar (Rockwell Intl. Science

Center)

Scientific Committee:

R. Agarwal (McDonnell Douglas); J. Ambrosiano (LLNL); G. Borghiotti (GWU);

J. Crowley (DARPA); A. Drobot (SAIC, McLean); J. Fratemico (SAIC, San Diego);

P. Goorjian (NASA Ames); M. Hafez (UC Davies); P. Joly (INRIA); J. Kong (MIT);

B. Mercier (Aerospatiale);R. Mittra (U. Illinois); K. Morgan (Swansea);

J.C. Nedelec (Ecole Polytechnique); J.P. Priou (MOD, France);

V. Rokhlin (Yale); T.K. Sarkar (Syracuse); B. Scheurer (CEA); J. Shang (WPAFB);

J. Ziolkowski (U. Arizona)

ABSTRACT DEADLINE : May 1st, 1993.

Proposed title and 2-3 pages extended abstract sould be sent to:

US and non-European participants: European participants:

R. Lohner J. Periaux

Office of Conferences and Institutes GAMNI

The George Washington University INRIA, Projet Menusi

2003 G Street, NW Domaine de Voluceau

Washington, DC 20052 Rocquencourt

B.P. 105

78153 Le Chesnay, Cede

FRANCE

FAX: 202-994-7400 FAX: 33-1-39.63.53.3

TEL: 202-994-7238 TEL: 33-1-39.63.54.82

e-mail: lohner@seas.gwu.edu e-mail: periaux@menusin.inria.fr

WORKSHOP TEST CASES

Instructions for retrieving the test-cases data:

Through anonymous ftp:

On your home machine "cd" to the directory where you want

the data to go; then enter the following:

ftp ftp.gwu.edu

at the prompt, enter:

anonymous

and then enter your e-mail address as the password. Then type:

cd pub/emworkshop

binary

prompt

mget *

quit

That should do it.

------------------------------

From: Trini Flores <flores@siam.org>

Date: Fri, 16 Apr 93 08:44:51 EST

**Subject: ACM-SIAM Symposium on Discrete Algorithms**

SODA 94 - - CALL FOR PAPERS

Fifth Annual ACM-SIAM Symposium on Discrete Algorithms

Sponsored by ACM-SIGACT and SIAM Activity Group on Discrete

Mathematics

January 23-25, 1994

Key Bridge Marriott Hotel

Arlington, Virginia

Organizer and Chair: Daniel D. Sleator, Carnegie-Mellon

University

The symposium is for those interested in the use, design, and

analysis of algorithms. Papers about algorithms (sequential,

parallel, distributed, randomized, etc.) in all areas of

application are invited.

Papers will be selected for presentation based on extended

abstracts. Authors wishing to submit a paper should send twelve

(12) copies of an extended abstract (not a full paper) by

Tuesday, July 6, 1993 to:

SIAM Conference Coordinator

3600 University City Science Center

Philadelphia, PA 19104-2688 USA

Abstracts should begin with a succinct statement of the problems

that are considered in the paper, the main results achieved, an

explanation of the significance of the work, and a comparison to

past research. This material should be easily understood by

nonspecialists. Technical development, directed toward the

specialist, should follow as appropriate. The entire extended

abstract must not exceed ten double-spaced typed pages (about

12,000 bytes). Abstracts that deviate significantly from these

guidelines risk rejection without consideration of their merits.

Abstracts must be received by the July 6 deadline otherwise

submissions will not be considered.

For additional information, contact SIAM.

Telephone: 215-382-9800

Fax: 215-386-7999

E-Mail: meetings@siam.org

------------------------------

From: Maurice.Meneguzzi@cerfacs.fr

Date: Thu, 15 Apr 1993 16:10:29 +0200

**Subject: Position at CERCA, Montreal**

CERCA 5160, boul. Decarie, bureau 400

CEntre de Recherche Montreal (Quebec) Canada H3X 2H9

en Calcul Applique Telephone:(514) 369-5224

Fax:(514) 369-3880

CEntre for Research on Computation and its Applications

JOB ANNOUNCEMENT:

RESEARCH PROFESSIONAL FOR LARGE SCALE PROBLEM SOLVING

CERCA invites applications for the position of "Research Professional"

in the area of large scale problem solving.

Candidates for the position, who must have a Master's or a doctoral

degree, are required to have training in the area of numerical methods and, in

particular, to be familiar with iterative methods, preconditioning techniques

and their application to the finite element method. Experience with parallel

computing and vectorization would be an asset.

The selected candidate will be expected to contribute to the refinement

of state-of-the-art techniques for the solution of large scale problems

associated with numerical simulation in fluid mechanics. He will also be

expected to cooperate in the operation of applied projects and in the

development of new methods.

CERCA is a nonprofit corporation. It was established in January 1992,

following an initiative from E'cole Polytechnique, Concordia University, McGill

University, and the Universite' de Montre'al, with an infrastructure grant from

the Ministe`re de l'Enseignement supe'rieur et de la Science du Quebec.

Currently, Alcan International, Bombardier-Canadair, CANMET (Energy, Mines and

Resources Canada), Environment Canada (AES, Dorval), GE Canada, GEC

Alsthom E'lectrome'canique, Hydro-Que'bec (Equipment Group), and Pratt &

Whitney Canada are regular industrial members of CERCA. In addition, Spar

Aerospace and Bio-Mega are associate industrial members.

Interested candidates should submit a detailed re'sume', and names,

addresses and phone numbers of three professional references, to Professor

Andre' Biron, Associate Director, CERCA, 5160 boul. De'carie, Suite 400,

Montre'al, Que'bec H3X 2H9, Canada. Applications received before May 14,

1993 will be assured consideration.

CERCA offers competitive salaries and fringe benefits to its employees,

and an excellent environment for professional growth. In accordance with

Canadian immigration requirements, priority will be given to citizens and

permanent residents of Canada.

This advertisement applies equally to women and men.

------------------------------

From: SIAM <tate@siam.org>

Date: Wed, 14 Apr 93 16:51:23 EST

**Subject: Contents, SIAM Math Analysis**

Contents, SIAM J. Math. Anal. July 1993

Convexity of Solutions to Some Elliptic Partial Differential Equations

Antonio Greco and Giovanni Porru

One-Phase Riemann Problem and Wave Interactions in Systems of Conservation

Laws of Mixed Type

Haitao Fan

Solitary Waves in a Two-Layer Fluid with Surface Tension

S.M. Sun and M. C. Shen

The Linear Shallow Water Theory: A Mathematical Justification

James A. Donaldson and Daniel A. Williams

Nonlinear Stability and Asymptotic Behavior of Shearing Motions of a Non-

Newtonian Fluid

John A. Nohel and Robert L. Pego

Antiplane Shearing Motions of a Visco-Plastic Solid

J.M. Greenberg and Anne Nouri

A Geometric Singular Perturbation Analysis of Detonation and Deflagration Waves

I. Gasser and P. Szmolyan

Travelling Waves for Mutualist Species

Konstantin Mischaikow and Vivian Hutson

Stability of Equilibria for a Class of Time-Reversible, Dn x 0(2)-Symmetric

Homogeneous Vector Fields

I-Heng McComb and Chjan Q. Lim

Unconditional Bases of Wavelets for Sobolev Spaces

Gustaf Gripenberg

Least Squares Approximation by Radial Functions

E. Quak, N. Sivakumar, and J. D. Ward

Superharmonic Solutions for Some Second-Order Differential Equations with

Singularities

Alessandro Fonda, Raul Manasevich, and Fabio Zanolin

Estimates for Large Deviations in Random Trigonometric Polynomials

George Benke and W.J. Hendricks

Selberg Integrals and Hypergeometric Functions Associated with Jack Polynomials

Jyoichi Kaneko

------------------------------

From: Arieh Iserles <A.Iserles@damtp.cambridge.ac.uk>

Date: Thu, 15 Apr 93 11:58:44 BST

**Subject: Contents, Acta Numerica**

Contents of Acta Numerica 1993:

Gene Allgower and Kurt Georg (Colorado State University):

"Continuation and path-following" ................................1

Carl de Boor (University of Wisconsin): "Multivariate splines" ..65

Jim Demmel (University of California at Berkeley), Mike Heath

(University of Illinois) and Henk van der Vorst (University of

Utrecht): "Parallel numerical linear algebra" .................111

J.L.M. van Dorsselear, J.F.B.M. Kraaijevanger and M.N. Spijker

(Rijksuniversiteit Leiden): "Linear stability analysis in the

numerical solution of initial value problems" .................199

Michel Fortin (Universite Laval): "Finite element solution of the

Navier--Stokes equations" ......................................239

Harry Yserentant (University of Tuebingen): "Old and new convergence

proofs for multigrid methods" ..................................285

------------------------------

End of NA Digest

**************************

-------