Subject: NA Digest, V. 93, # 16
NA Digest Sunday, April 18, 1993 Volume 93 : Issue 16
Today's Editor:
Cleve Moler
The MathWorks, Inc.
moler@mathworks.com
Today's Topics:
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
Submissions for NA Digest:
Mail to na.digest@na-net.ornl.gov.
Information about NA-NET:
Mail to na.help@na-net.ornl.gov.
-------------------------------------------------------
From: Glenn R. Luecke
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
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
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
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
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
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
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
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
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
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
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
**************************
-------