NA Digest Sunday, October 6, 1991 Volume 91 : Issue 40

Today's Editor: Cleve Moler

Today's Topics:

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: Bill Rosener <rosener@cs.utk.edu>
Date: Wed, 2 Oct 91 15:03:34 -0400
Subject: NA-NET Whitepages

NA-NET Whitepages

This is just a reminder that the NA-NET supports a whitepages feature.
It allows users of the NA-NET to find out more information about other
members. We would like to encourage you to join if you have not done
so. Like any directory service, it will only work if everybody takes
a few moments to join, and then as necessary update their entry.

For more information on the whitepages send mail to
"na.help@na-net.ornl.gov"

Below are a few examples:

1). To find out more information on all people whose last name
begins with "van".

mail na.whois@na-net.ornl.gov
Subject: van

Null body


2). To find all members who are interested in "parallel".

mail na.whois@na-net.ornl.gov
Subject: blank

Keyword: parallel


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

From: Gerald W. Hedstrom <hedstrom@physics.llnl.gov>
Date: Wed, 2 Oct 91 13:43:03 PDT
Subject: Journal of Computational Physics Editors Resign

The editors of the Journal of Computational Physics have
resigned effective October 1, 1991. Authors and referees should
send correspondence concerning manuscripts to Academic Press, Inc.,
1250 Sixth Avenue, San Diego, CA 92101.

Gerald Hedstrom
Lawrence Livermore National Laboratory


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

From: Brian Morris <brian@panther.math.oxy.edu>
Date: Sun, 29 Sep 91 17:44:46 PDT
Subject: Change of Address for Brian C. Morris

For the academic year 91-92, I will be at
Occidental College, Department of
Mathematics, Los Angeles, CA 90041
213-259-2504

My e-mail adress there is:
brian@oxy.edu

-- Brian Morris


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

From: Bill Venables <wvenable@stats.adelaide.edu.au>
Date: Mon, 30 Sep 1991 15:25:44 +0930
Subject: Re: Journal Refereeing is Breaking Down

In NA Digest, V. 91, # 39 Steve Vavasis put forward a suggestion for
overcoming the delays in the refereeing system, namely that

"SIAM adopt the following policy. SIAM would maintain a list of
people who carry their share in terms of refereeing, and of
people who repeatedly decline to referee SIAM papers. Anyone
in the latter category would not be allowed to publish in SIAM
journals."

While I have some sympathy with this suggestion on rough equity grounds
(having suffered from the problem myself, and on occasion, having contributed
to it), I think it would be a dangerous policy for SIAM to persue. It would
give the unfortunate impression that excellence of research was not the sole
criterion for acceptance for publication in SIAM journals, but that somehow
publication was in part a reward for services rendered.

It would also encourage a rather slipshod approach to refereeing and a large
increase in arbitrariness. In a sense refereeing work is so important, so
difficult and so thankless it has to be both confidential and voluntary.

Bill Venables, Dept. of Statistics, | Email: venables@stats.adelaide.edu.au
Univ. of Adelaide, South Australia.| Phone: +61 8 228 5412


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

From: William C. Gear <cwg@research.nj.nec.com>
Date: Mon, 30 Sep 91 09:04:42 EDT
Subject: Refereeing Delays

Steve Vavasis (vavasis@cs.cornell.edu) raised the issue of the increase
in refereeing delays in a recent NA newsletter. He suggested that SIAM,
at least, might do something about it by maintaining lists of those in
the community who are responsive to requests to referee and those who
consistently decline, and then decline to publish papers from those
who consistently decline (a sort-of Santa Clause's nauty and nice list).
While I would love to find some solutions to the refereeing delay problem
(and so would all of the SIAM staff) I am hesitant to use such a procedure.
There are, for example, some authors who write excellent papers who are
appalling referees, not from a refusal to do it but froma point of view
of standards, either unreasonably harsh or the other extreme. On the other
hand, there are some people who contribute enormously to our profession
in various hidden ways, and one such way is by conscientious refereeing,
and yet do not generate much in the way of outstanding papers. (Incidentally,
I would like to find a way to reward these people more than our profession
does.)

In other words, I think we should be careful about linking publication with
the amount of refereeing performed: it has the potential to lower the standards
of the journals. (Possibly we could use the information that people had
consistently declined to be referees or been extremely tardy to push their
papers back in the priority queue which determines the publication date
once the paper has been accepted--currently papers are alloacated to that
queue on a first-come-first-served basis. However, I would be concerned
about the consistent and correct collection of the data, and how we would
balance the refereeing performance against other activities. Should we
expect the same response to refereeing requests from a person who is
active in a range of professional activites such as running SIAGs or
refereeing for non-SIAM journals as from someone who does little else?
In fact, the old adage that when you want something done you go to the
busiest person on the block usually works for refereeing!)

The major problem of delays seems to be one of responsibility. If a
referee knows he or she can't do the job very quickly, the paper
should be returned immediately (better yet, inform the editor by email).
If the paper is kept, it should be refereed in month. Allowing for mail delays
that will give around a two month turn around to the author. I have heard
many people say "Well, I have a stack of things to referee, and it goes on
the bottom in its priotity order." I don't think that is being responsible.
If you have a stack that is going to cause a month or two delay, decline
to referee another item AND clear the stack. On the average, everybody
should be refereeing about twice as many papers as they write in the case
of SIAM journals, since the acceptance rate is around 50%.

I would love to get comments from the community and ideas how we can
improve the situation. A discussion in this forum would be good: it would
be helpful to me if you would cc me at cwg@research.nj.nec.com

Bill Gear

Vice President of Publications,
Society for Industrial and Applied Mathematics

UUCP: princeton!nec!cwg
PHONE: (609) 951-2700
FAX: (609) 951-2483


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

From: Carl Taswell <taswell@na-net.stanford.edu>
Date: Wed, 2 Oct 91 15:59:04 PDT
Subject: Journal Refereeing

In response to Steve Vavasis' submission entitled "Journal Refereeing
is Breaking Down", I would like to offer a few questions for consideration.

How should the NA community ensure not only that papers are refereed
in a timely manner by those qualified as sufficiently knowledgeable
experts in their fields, but also that when they do referee papers,
that they do so in a competent and responsible (ie, accountable) manner?

Should incentives be carrots or sticks? What about the possibility of
allowing referees the privilege of publishing a brief paragraph
summarizing their analyses of the submitted paper?

Should the review process be made less time consuming for the
referees? This could be done in a number of ways such as requiring the
submission of detailed derivations of proofs and formulas, and
detailed algorithms with actual computer code, not intended for
publication necessarily, but rather for quick and easy verification by
the referee. This verification would guarantee that proposed methods
will in fact be reproducible by other investigators.

I'm sure there are other possibilities that could also be considered...


Carl Taswell
taswell@na-net.stanford.edu


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

From: Yair Censor <RSMA403%HAIFAUVM@TAUNIVM.TAU.AC.IL>
Date: Thu, 03 Oct 91 09:54:13 IST
Subject: Journal Refereeing is Breaking Down

In his recent note :" Journal Refereeing is Breaking
Down" (NA Digest, Sunday, September 29, 1991, Volume 91: Issue 39),
Steve Vavasis suggests that: "....people who repetedly decline to referee
SIAM papers..... would not be allowed to publish in SIAM journals".
While the problem he points at is indeed a serious one, his solution
is totally unacceptable. Without even entering the many "practical"
difficulties of implementing such a policy, it must be rejected up front
because of its contradiction to the very nature of the freedom of flow
of information in science.

--yair censor
rsma403@haifauvm.bitnet


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

From: Fred Kus <FRED@SSCvax.CIS.McMaster.CA>
Date: Fri, 4 Oct 1991 14:13 EDT
Subject: Fortran Code for NETFLO

Hi,

I would greatly appreciate any info about the NETFLO algorithm, a linear
programming routine for Network Programming. Is there a Fortran
implementation available, preferably in the public domain ?
Thanks.

Fred

Fred W. Kus INTERNET: fred@SSCvax.CIS.McMaster.CA
Computing & Information BITNET: fred@MCMASTER.BITNET
Services PHONE: (416) 525-9140 ext.4160
Mcmaster University FAX (416) 528-3773
Hamilton, Canada L8S 4K1 |



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

From: Jerry Tiemann <tiemann@snoopy.crd.ge.com>
Date: Tue, 1 Oct 91 14:46:48 EDT
Subject: Numerical Recipes Algorithms

I have found that the Nelder-Meade simplex algorithm for fitting data to a
superposition of parametrized component curves has trouble even when the
target data is a "perfect" superposition of such components. In the
simplest cxase, where the constant factors are considered to be part of the
parameter space, there are often too many local minima to get a reliable
fit. If one separates out these constants (as is done in the MATLAB Demo),
and fits these in a separate linear algebra step, the problem is less
severe, but it still often converges prematurely. The problem appears to be
that the Nelder-Meade step, in finding a local minimum IN THE CONTEXT OF
THE CURRENT linear coefficients, finds an optimum in which these constants
are indeed optimal. Hence the problem. To fix this, I have tried two
variants: 1) Don't use the best fit coefficients; use a linear combination
(either accelerated or damped) of the old coefficients and the best fit
ones. 2) Use a different cost function for the Nelder-Meade step than the
2-norm of the error. (i.e. use absolute value or 4-norm or something else).
The idea here is that by using a different error surface in the two steps of
each improvement cycle, the chances of falling into a minimum that is at
the exact same location in parameter space is smaller. This second
approach appears to help quite a bit, but even with this improvement, the
problem sometimes occurs.


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

From: Iain Duff <ISD%IBM-B.RUTHERFORD.AC.UK@VTVM2.CC.VT.EDU>
Date: Fri, 04 Oct 91 14:49:08 BST
Subject: Harwell Subroutine Library

HARWELL SUBROUTINE LIBRARY

It is now over a year since Nick Gould, John Reid, Jennifer Scott and myself
moved over from Harwell (now part of AEA Technology) to the Atlas Centre at
the Rutherford Appleton Laboratory. Since then I and my colleagues have
frequently been asked about the status of the Harwell Subroutine Library and
our relationship with it.

I am very pleased to relate that a new agreement has recently been reached
between the Science and Engineering Research Council of Great Britain (SERC)
and AEA Technology which will ensure the continued development
of the Library and our involvement with it.

The main points of the agreement are that the Harwell Subroutine Library
will continue to be developed and supported by us (the Numerical Analysis
Group at RAL) and will be marketed by AEA Technology with revenues shared
between SERC and AEA Technology. The Library will be updated on a regular
basis, the current release being Release 10 as presently marketed by Harwell.
For the immediate future, the continuity will be emphasized by retaining the
name Harwell Subroutine Library. Current plans are to issue the next Release
in just over one year's time. The agreement with the Numerical Algorithm
Group (NAG Inc) to market the subset entitled "The Harwell Sparse Matrix
Library" will continue, and we expect to update that Library at the end of 1992.

The agreement with Harwell explicitly encourages joint development work with
third parties and we can provide collaborators with royalty free licences
for existing software to assist with our joint work.

Users wishing to mount and use HSL on machines in their own University or
Laboratory can obtain licences through Harwell (Elizabeth Thick, Bldg 424.4,
AEA Technology, Harwell Laboratory, Didcot, Oxon OX11 0RA. Tel: +44-235-432688)
to whom matters relating to administrative Library details should be addressed.
Comments on library contents, collaborative research, and technical issues
can be addressed (preferably by email) to any of the Numerical Analysis
Group ...

Iain Duff isd@uk.ac.rl.ib
Nick Gould nimg@uk.ac.rl.ib
John Reid jkr@uk.ac.rl.ib
Jennifer Scott sct@uk.ac.rl.ib

Iain S Duff
Atlas Centre
Rutherford Appleton Laboratory
Tel: +44-235-445803
Fax: +44-235-446626

PS .. Please note that the Fax number for all members of the Numerical
Analysis Group is now +44-235-446626


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

From: Michael Overton <overton@OVERTON.CS.NYU.EDU>
Date: Wed, 2 Oct 91 08:37:23 -0400
Subject: Tech Reports Available on Anonymous ftp

NYU Computer Science Department Technical Reports are now available by
anonymous ftp. The numerical analysis faculty members at NYU participating
in this service are myself, Anne Greenbaum and Olof Widlund. We have
chosen to make dvi files available, in compressed form. Last week's digest
contribution by Pete Stewart outlines various alternative options that
institutions might use; our choice use a somewhat different format to Pete's
but is consistent with his overall suggestions. If our conventions were
followed widely, a potential user would not need to know anything beyond
the general form of the conventions and an internet address of the form
department.university.edu, the emerging standard for e-mail address formats.

In order to find out what is available, do
ftp cs.nyu.edu
and login as anonymous, providing identification as password. Then do
cd pub/local
ls
and you will see names of faculty, including overton, greenbau (note the
Unix truncation), widlund. To find about my files, for example, do
cd overton
ls
and you will see what files are available. All of them end in dvi.Z,
indicating a dvi file in compressed form, except the index file, called
index.readme....to get a copy of any of these files, simply do
get file
To quit, do
quit
Then you can examine the index.readme file on your own machine, or process
a report file by executing "uncompress xxxx.dvi.Z" and then your favorite
dvi processor, e.g. dvitool xxxx or dvips xxxx. The index file gives a lot
of information, including journal submission/publication details, and so
could be useful for bibliographic searches.

Alternatively, you can access reports by report number by executing
cd /
cd pub/tech-reports
ls
and "getting" the file name with the technical report number you want.
The latter files are links to the former.

Challenge to the community: implement a standard along these lines, and
help solve the national landfill problems by not mailing unsolicited paper
copies of reports to huge mailing lists.


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

From: Nick Trefethen <lnt@cs.cornell.edu>
Date: Wed, 2 Oct 91 17:46:48 -0400
Subject: Bibliography on Non-normal Matrices and Pseudospectra

A bibliography on the subject of "pseudo-eigenvalues" of
non-normal matrices and operators is available by anonymous
ftp to ftp.cs.cornell.edu. (See Pete Stewart's recent NA-NET
contribution about anonymous ftp.) The file is called
/pub/trefethen/pseudo.tex (or pseudo.dvi, or pseudo.ps) and
contains also a brief summary of this subject.

Nick Trefethen
Dept. of Computer Science
Cornell University
lnt@cs.cornell.edu


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

From: Esmond Ng <esmond@msr.epm.ornl.gov>
Date: Thu, 3 Oct 91 14:54:03 -0400
Subject: Second Announcement of the Tenth Parallel Circus

Second announcement of the Tenth Parallel Circus -

The Mathematical Sciences Section and the Engineering Physics and
Mathematics Division at the Oak Ridge National Laboratory, Oak Ridge,
Tennessee will be hosting the Tenth Parallel Circus in Oak Ridge on
Friday and Saturday, October 25-26, 1991. For details, see NA Digest
Vol. 91, Issue 33 (August 18, 1991), or send e-mail to
workshop@msr.epm.ornl.gov.

Limited financial support may be available to eligible graduate
students. For details, see NA Digest Vol. 91, Issue 39 (September 29,
1991), or send e-mail to na.golub@na-net.ornl.gov.


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

From: John W. Petro <petro@gw.wmich.edu>
Date: Sun, 06 Oct 91 08:46:43 IST
Subject: Symposium on Matrix Analysis and Applications

Symposium on Matrix Analysis and Applications

Western Michigan University, Friday and Saturday, 11-12 October 1991
Dedicated to the Memory of Erik A. Schreiner

All lectures will be in the Commons Room, Sixth Floor, Everett Tower,
Western Michigan University.

For more information and a detailed schedule, please contact John W. Petro.
E-mail: petro@gw.wmich.edu
Phone: 616-387-4551

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

From: University of Minnesota <glenys@cs.umn.edu>
Date: Tue, 1 Oct 1991 14:26:42 -0500
Subject: Openings for Postdocs and R.A.'s at Minnesota

University of Minnesota
Department of Computer Science

Openings for postdocs and graduate research assistants The Scientific
Computing Group in the Department of Computer Science and the Army High
Performance Computing Research Center have openings for postdocs and
graduate research assistants in various aspects of Numerical Analysis and
the development of Parallel Numerical and Nonnumerical Algorithms. Our
research activities include: numerical methods for real-time solution of
differential-algebraic equations, parallel algorithms in sparse matrix
computations and large-scale optimization and control, as well as
scalability analysis of parallel systems. The scientific computing faculty
in the Department of Computer Science has expanded substantially last year
and includes: Dan Boley, Anthony Chronopoulos, David Fox, Vipin Kumar,
Haesun Park, Linda Petzold, Ben Rosen, Youcef Saad, and Ahmed Sameh.

The University of Minnesota`s Army High Performance Computing Research
Center (AHPCRC) complements the activities of this group. AHPCRC's
interdisciplinary research program focuses on the high performance
computational aspects of large-scale applications in science and
engineering. State-of-the-art computing facilities such as the newest
Thinking Machine Corp. CM-5, the Cray 2/4-512 and Cray X-MP/4-6 are
available for research throuh AHPCRC and the Minnesota Supercomputer
Center(MSC). The Department of Computer Science, the Minnesota
Supercomputer Institute, the Institute of Mathematics and its Applications,
AHPCRC and the Geometry Center all have active lecture series/visiting
programs in various aspects of scientific computing, and there are great
opportunities for interaction with prominent researchers in scientific,
medical, and engineering disciplines.

The University of Minnesota is an equal opportunity educator and employer.

Please send your questions or applications to:
Linda Petzold
Department of Computer Science
4-192 EE/CS, 200 Union Street S.E.
University of Minnesota
Minneapolis, MN 55455
email: petzold@cs.umn.edu


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

From: Graeme Fairweather <MAT110@UKCC.uky.edu>
Date: Fri, 04 Oct 91 13:26:08 EDT
Subject: CFD Position at the University of Kentucky

The University of Kentucky
Department of Mathematics

Applications are invited for a junior/senior position in
computational fluid dynamics beginning August 16, 1992. The
successful candidate must demonstrate expertise in finite
element applications, spectral methods or grid generation.
Please send a curriculum vitae, sample publications, and
have three or four letters of recommendation sent to James
H. Wells, Chairman. The University of Kentucky is an equal
opportunity, affirmative action employer. Applications of
women and minorities are especially encouraged.


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

From: William Hager <hager@math.ufl.edu>
Date: Mon, 30 Sep 91 17:01:55 EDT
Subject: New Journal in Computational Optimization

A new journal entitled Computational Optimization and Applications
is being established by Kluwer Academic Publications. A copy of the aims
and scope for the journal and the editorial board appear below. The author
instructions can be obtained from either of the following addresses:

hager@math.ufl.edu Karen@world.std.com

William W. Hager Karen S. Cullen
Department of Mathematics Kluwer Academic Publishers
201 Walker Hall Editorial Office
University of Florida 101 Philip Drive
Gainesville, FL 32611 Norwell, MA 02061

CALL FOR PAPERS

COMPUTATIONAL OPTIMIZATION AND APPLICATIONS,

An International Journal

Aims and Scope

Computational Optimization and Applications is a peer
reviewed journal that is committed to timely publication of
research and tutorial papers on the analysis and development
of computational algorithms and modeling technology for
optimization. Algorithms either for general classes of
optimization problems, or for more specific applied problems
are of interest. Stochastic algorithms as well as deter-
ministic algorithms will be considered. Papers that can
provide theoretical analysis along with carefully designed
computational experiments are particularly welcome. Topics
of interest include, but are not limited to the following:

o Large Scale Optimization
o Unconstrained Optimization
o Linear Programming
o Quadratic Programming, Complementarity Problems,
and Variational Inequalities
o Constrained Optimization
o Nondifferentiable Optimization
o Integer Programming
o Combinatorial Optimization
o Stochastic Optimization
o Multiobjective Optimization
o Network Optimization
o Complexity Theory
o Approximations and Error Analysis
o Parametric Programming and Sensitivity Analysis
o Parallel Computing, Distributed Computing, and Vector Processing
o Software, Benchmarks, and Numerical Experimentation and Comparisons
o Modeling Languages and Systems for Optimization
o Automatic Differentiation
o Applications in Engineering, Finance, Optimal Control, Optimal Design,
Operations Research, Transportation, Economics, Communications,
Manufacturing, and Management Science


EDITORIAL BOARD


Jean Abadie (France)
Neculai Andrei (Romania)
Achim Bachem (Germany)
Egon Balas (Carnegie Mellon)
Jaime Barcelo (Spain)
Dimitri P. Bertsekas (MIT)
John R. Birge (Michigan)
Johannes J. Bisschop (Holland)
Jens Clausen (Denmark)
Thomas M. Cook (American Airlines Decision Technologies)
George B. Dantzig (Stanford)
Michel C. Delfour (Canada)
Ronald S. Dembo (Canada)
Vladimir F. Demyanov (USSR)
Gianni Di Pillo (Italy)
Jaroslav Dolezal (Czechoslovakia)
Asen L. Dontchev (Bulgaria)
Joseph C. Dunn (NCSU)
Anthony V. Fiacco (George Washington)
Masao Fukushima (Japan)
Jean-Louis Goffin (McGill)
William W. Hager (Florida)
Patrick T. Harker (Pennsylvania)
Donald W. Hearn (Florida)
David R. Heltne (Shell Development)
Karla L. Hoffman (George Mason)
Masao Iri (Japan)
Johannes Jahn (Germany)
Joaquim J. Judice (Portugal)
Jeffery L. Kennington (Southern Methodist)
Hiroshi Konno (Japan)
Leon Lasdon (Texas)
Irena Lasiecka (Virginia)
Siriphong Lawphongpanich (Naval Postgraduate School)
P. O. Lindberg (Sweden)
Thomas L. Magnanti (MIT)
Istvan Maros (Hungarian Academy of Sciences)
Gautam Mitra (England)
Sang Nguyen (Canada)
Jong-Shi Pang (Johns Hopkins)
Panos Pardalos (Penn State)
Elijah Polak (Berkeley)
Boris T. Polyak (USSR)
Aubrey B. Poore (Colorado State)
Rouben Rostamian (Maryland-Catonsville)
Ekkehard W. Sachs (Germany)
Naum Z. Shor (USSR)
Jan Sokolowski (Poland)
Monique Guignard-Spielberg (Pennsylvania)
Gilbert Strang (MIT)
Stavros A. Zenios (Pennsylvania)


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

From: Iain Duff <ISD%IBM-B.RUTHERFORD.AC.UK@VTVM2.CC.VT.EDU>
Date: Tue, 01 Oct 91 17:49:22 BST
Subject: Ron Mitchell Issue of IMA Journal of Numerical Analysis

IMA JOURNAL OF NUMERICAL ANALYSIS - Volume 11, Number 4

This issue comprises papers dedicated in honour of the 70th birthday of
Professor A (Ron) Mitchell. For reasons of space, all papers so dedicated
could not appear in this issue. The remaining papers will appear in
following issues of IMAJNA.


J M Sanz-Serna and A new class of results for the algebraic equations
D F Griffiths of implicit Runge-Kutta processes.

D J Higham Global error versus tolerance for explicit
Runge-Kutta methods.

G A Watson An algorithm for optimal l scaling of matrices.
2

I Christie and An exact Riemann solver for a fluidized bed model.
C Palencia

Y Tourigny Optimal H 1 estimates for two time-discrete Galerkin
approximations of a nonlinear Schrodinger equation.

G Fairweather and A box method for a nonlinear equation of population
J C Lopez-Marcos dynamics.

M J Ablowitz, Dynamics of semi-discretizations of the defocusing
B M Herbst and nonlinear Schrodinger equation.
J A C Weideman

S W Schoombie Stability properties and spurious period two
solutions in a numerical scheme for a reaction-
diffusion equation with nonlinear diffusion.

J W Barrett, Finite element approximation of a rigid punch
R Chakrabarti and indenting a membrane.
C M Elliott

T Diego, S McKee A Hermite-type collocation method for the solution
and T Tang of an integral equation with a logarithmic singular
kernel.


The subscription rates for the IMA Journal of Numerical Analysis for 1992
have been set by OUP at 120 for EEC and UK and $260 for North America
and the rest of the world. The price to members of the Institute of
Mathematics and its Applications is 48 per annum.


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

From: Robert G. Voigt <rgv@icase.edu>
Date: Tue, 1 Oct 91 13:29:59 -0400
Subject: Scalable High Performance Computing Conference

SHPCC '92 Call for Papers

Scalable High Performance Computing Conference
April 26-29, 1992, Williamsburg Hilton, Williamsburg, VA.
Sponsored by IEEE

The Scalable High Performance Computing Conference (SHPCC) is the
successor to the Distributed Memory Computing Conference series. We
define a scalable high performance architecture as an architecture that
is likely to be capable of delivering a teraflop sustained performance
in the relatively near future. SHPCC will focus on software being
developed to make it possible to effectively exploit the coming generations
of these architectures and on applications that require teraflop
computational rates.

We invite papers from researchers who are developing or evaluating
software tools such as compilers, programming environments, and debuggers
along with tools to monitor and tune performance. Papers are also
solicited from those who have developed and analyzed multiprocessor
applications codes. We plan to place particular emphasis on two
general application areas. The first is that of computational
fluid dynamics and its uses in problems that arise in the design of
aircraft, spacecraft and automobiles. The second application area
consists of molecular dynamics and its uses in the pharmaceutical
and chemical industries.

The conference will begin with one day of tutorials. The two and one
half day technical program that follows will consist of four invited talks,
two parallel tracts of half hour talks, and a small poster session.
We invite researchers to submit a four page extended abstract in hard
copy, LaTeX, or Troff to the address below by November 25, 1991.
The abstracts will be carefully reviewed by the program committee and
the authors will be notified by early January, 1992. The proceedings
will be published by IEEE and will be available at the time of the
conference. Final papers will be due approximately, March 1, 1992.

Ms. Emily Todd,
ICASE,
Mail Stop 132-C,
NASA Langley Res. Ctr.,
Hampton, VA 23665
email: emily@icase.edu

General Chair: Rober Voigt, ICASE

Program Committee:

Joel Saltz, ICASE, Chair
Joseph Brandenburg, Intel Corp.
Geoffrey Fox, Syracuse Univ.
Dennis Gannon, Indiana Univ.
Andrew Grimshaw, Univ. of Virginia
Micheal Heath, Univ. of Illinois
Lennart Johnnson, Harvard Univ. & Thinking Machines Corp.
Robert Martino, National Institutes of Health
Paul Messina, California Institute of Technology
Dan Reed, Univ. of Illinois
Manuel Salas, NASA Langley Research Center
Horst Simon, Computer Sciences Corp.


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

From: Tony Chan <chan@math.ucla.edu>
Date: Tue, 1 Oct 91 11:48:43 -0700
Subject: Hestenes Memorial Symposium

ANNOUNCEMENT
Magnus Hestenes Memorial Symposium
Department of Mathematics
University of California, Los Angeles


The Department of Mathematics of UCLA will host a memorial symposium
in honor of Professor Magnus Hestenes on Saturday, November 23, 1991.
Professor Hestenes passed away on May 31, 1991. He had been a faculty
member in the UCLA's Math Dept. since 1947 until his retirement in 1973.
At this symposium, the speakers will pay both personal and technical
tribute to Professor Hestenes. He contributed to many areas of
mathematics, including conjugate gradient methods, calculus of variations,
optimization theory and optimal control. He also played an active
role in leading UCLA's Department of Mathematics and the
National Bureau of Standards Institute for Numerical Analysis at UCLA.

Symposium speakers will include:
A.V. Balakrishnan (UCLA),
Len Berkovitz (Purdue),
Roland Glowinski (U. of Houston),
Gene Golub (Stanford),
David Hestenes (Arizona State U.),
John Hestenes (NSF),
Ed Landesman (U.C. Santa Cruz),
Richard Tapia (Rice),
John Todd (Caltech).

All interested parties are welcome to attend this symposium.
There will be no registration fees.
A banquet will be organized the same evening at the UCLA Faculty Center
(cost approximately $32.00 per person all inclusive).
Please register and reserve seats for the banquet before November 8, 1991.
For registration and more information please contact
Babette Dalton at (213) 825-9036; e-mail: babette@math.ucla.edu.


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

From: Tom Manteuffel <tmanteuf@copper.Denver.Colorado.EDU>
Date: Fri, 4 Oct 91 09:01:01
Subject: Copper Mountain Conference on Iterative Methods

CALL FOR PAPERS
Copper Mountain Conference
on
Iterative Methods

Sponsored by DOE
In Cooperation with the SIAM Activity Group on Linear Algebra

COPPER MOUNTAIN, COLORADO
April 9-14, 1992

Special Features:

Tutorial on Polynomial Iterative Methods
Student Paper Competition
Preliminary Proceedings
Special Journal Issues in SISSC and SIMAX

Topics of Emphasis:

Nonsymmetric Linear Systems
Nonlinear Systems
Applications on Advanced Architectures
Equivalent Preconditioning

Conference Deadlines:

Abstracts Jan 1, 1992
Student Papers Jan 1, 1992
Papers for Preliminary Proceedings Feb 15, 1992
Lodging Reservations March 1, 1992
Early Registration March 1, 1992

For Complete Information send email to:

cm-info@copper.denver.colorado.edu


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

From: Tom Manteuffel <tmanteuf@copper.Denver.Colorado.EDU>
Date: Fri, 4 Oct 91 09:01:01
Subject: Tutorial on Polynomial Iterative Methods

Tutorial on
Polynomial Iterative Methods

Denver, Colorado
April 7-8, 1992

Purpose:

This Tutorial is designed to provide a basic understanding of the
fundamentals of preconditioned polynomial iterative methods for the
solution of large sparse linear systems.

Format:

The Tutorial will be structured around a series of lectures and
computer workshops.

Tutorial Topics:

Basic Polynomial Iterative Methods,
Chebychev-like Methods,
Conjugate Gradient-like Methods,
Basic Preconditioning - Matrix Splitting,
Iterative Methods for Nonsymetric Systems,
Preconditioning by Incomplete Decomposition,
Equivalent Preconditioning

Program Staff:

Steve Ashby, Lawrence Livemore National Laboratories
Tom Manteuffel, University of Colorado at Denver
Paul Saylor, University of Illinois

For Complete Information send email to:

cm-info@copper.denver.colorado.edu


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

From: Gerald Richter <richter@cs.rutgers.edu
Date: Fri, 4 Oct 91 15:50:20 EDT
Subject: IMACS Conference on PDE's

FIRST ANNOUNCEMENT AND CALL FOR PAPERS
7TH. IMACS INTERNATIONAL CONFERENCE ON
COMPUTER METHODS FOR PARTIAL DIFFERENTIAL EQUATIONS
formerly known as the "Lehigh Conferences"
June 22-24, 1992, Rutgers University, New Brunswick, NJ 08903, USA

Manuscripts, proposals to organize sessions, and suggestions for
keynote speakers are solicited on all aspects of the numerical
solution of partial differential equations. While special emphasis
will be placed on those aspects that are related to applications of
new computing architectures, and those related to computational wave
propagation and fluid dynamics, contributions to other areas will also
be welcome.

Deadlines: -Submission of manuscripts (full texts or one page
comprehensive abstracts) and proposals to organize sessions:
As soon as possible.

-Notification of acceptance of manuscript: February 1, 1992

-Final manuscripts due: April 15, 1992

Conference Committee (provisional):
W. F. Ames (Georgia Tech.); M. J. Buckingham (Scripps Inst.);
J. Flaherty (RPI); M. Fritts (SAIC); A. Gerasoulis (Rutgers); D. Knight
(Rutgers); J. R. Rice (Purdue); G. Richter (Rutgers); R. Temam
(Univ. de Paris); J. Verwer (CWI-Amsterdam); R. Vichnevetsky
(Rutgers - Chairman)

As was the case for preceding IMACS PDE Conferences, Proceedings
containing all the accepted papers will be published.

Send Manuscripts (in duplicate), proposals to organize a session (or
sessions) to the Conference Committee, c/o:

Karen Hahn
IMACS PDE-7 Conference Secretary
Rutgers University
Department of Computer Science
New Brunswick, NJ 08903, USA
Fax: 908-932-5530
e-mail: imacs@cs.rutgers.edu

General correspondence and requests for information may be obtained
from the Conference secretariat at the above address or by contacting
members of the organizing committee directly.


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

From: Graeme Fairweather <MAT110@UKCC.uky.edu>
Date: Sat, 05 Oct 91 09:24:41 EDT
Subject: Mini-Conference on Domain Decomposition Methods

Mini-Conference on Domain Decomposition Methods
Preliminary Announcement

November 15 -16, 1991
University of Kentucky
Lexington, Kentucky

A mini-conference on Domain Decomposition Methods will be held in the
Center for Computational Sciences at the University of Kentucky,
Lexington, Kentucky. The conference will begin at 1 p.m. on Friday,
November 15 and end in the late afternoon of Saturday, November 16.

Invited speakers will include:

Tony F. Chan (UCLA)
Max Dryja (University of Warsaw/University of Kentucky)
Howard Elman (University of Maryland)
William D. Gropp (Argonne National Laboratory)
David E. Keyes (Yale University)
Barry F. Smith (Argonne National Laboratory)
Olof B. Widlund (NYU)
Bernard Bialecki (University of Kentucky)
X.-C. Cai (University of Kentucky)

Some contributed talks may also be scheduled.

To obtain further information, contact

Graeme Fairweather
Department of Mathematics
University of Kentucky
Lexington, Kentucky 40506

Telephone: 606-257-2326
E-mail: mat110@ukcc.uky.edu


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

From: Marcin Paprzycki <m_paprzycki@utpb.pb.utexas.edu>
Date: Sat, 5 Oct 1991 10:47:08 CDT
Subject: Permian Basin Supercomputing Conference

Call for Papers
PERMIAN BASIN SUPERCOMPUTING CONFERENCE 1992
March 13-15

The University of Texas of the Permian Basin
Odessa, Texas

Organized by: Marcin Paprzycki
Advisors: Firooz Khosraviyani, Mogens Melander

In cooperation with: SIAM Supercomputing Group

The purpose of the conference is to bring together researchers in
all fields of supercomputing, high performance computing and
parallel computing for an effective exchange of ideas and
discussion of recent developments and future directions of
research.

Sponsored by:
Center for High Performance Computing, Austin
Cray Research Inc.
DEC MPP 12000 Group
Intel Corporation
University of Texas of the Permian Basin

Invited Speakers:
Chris Bischof, Jurio Diaz, Ian Gladwell, David Kincaid,
George MacMechan, Robert Plemmons, Danny Sorensen,
Robert van de Geijn.

Submit three copies of detailed abstract (not to exceed 1000 words) by

December 20, 1992

to:
Marcin Paprzycki, Conference Chairman
Department of Mathematics and Computer Science
University of Texas of the Permian Basin
Odessa, TX 79762

Phone: (915) 367-2244
Fax: (915) 367-2115
E-mail: m_paprzycki@utpb.pb.utexas.edu
m_paprzycki@utpb.bitnet


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

End of NA Digest

**************************
-------