NA Digest Sunday, January 28, 1990 Volume 90 : Issue 04

Today's Editor: Cleve Moler

Today's Topics:

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


From: Will Light <maa007%central1.lancaster.ac.uk@nsfnet-relay.ac.uk>
Date: Tue, 23 Jan 90 15:20:47 GMT
Subject: SERC Numerical Analysis Summer School

SERC Numerical Analysis Summer School
Lancaster University
15 July - 3 August, 1990

The fourth SERC sponsored Numerical Analysis Summer School is to be held at
Lancaster next summer. The meeting will comprise three one-week modules each of
which can `stand alone' though it is hoped that many participants will stay for
longer periods. SERC will contribute towards the subsistence costs of a limited
number of participants each week. The meeting is open to all interested
researchers, but only those currently employed in the UK will be eligible for
SERC support.

In each week there will be three five-lecture courses given by the invited
instructors, additional activities organised by the local expert and a programme
of current research seminars given by participants. The outline programme for
the meeting is:

Non-linear Partial Differential Equations (15-20 July)

A. Quarteroni (Pavia) Spectral Methods
R. Nochetto (Maryland) Finite element methods for free and moving boundary
problems
L.B. Wahlbin (Cornell) Finite element methods for evolution equations
C. Elliott (Sussex) Local expert

Dynamical Systems and Bifurcation (22-27 July)

W.J. Beyn (Konstanz) Numerical methods for Dynamical Systems
J. Sanz-Serna (Valladolid) Nonlinear Stability
W.C. Rheinboldt (Pittsburg) Differential Equations on Manifolds
A. Spence (Bath) Local Expert

Multivariate Approximation (29 July - 3 August)

C. Chui(Texas A&M) Multivariate Splines
N. Dyn (Tel-Aviv) Computer-aided geometric design
M. Powell (Cambridge) Radial basis functions
J. Gregory (Brunel) Local expert

The lectures will take place from Monday to Friday each week and there will be
a substantial amount of time available for research and discussion with the
assembled experts. The aim of the specialist courses is to enable people with
initially only a nodding acquaintance with a topic to concentrate on it and end
up feeling quite competent in it, thus broadening their research horizons. It
is hoped that a participant with a good initial knowledge may be enabled to
work on significant problems in the area.

It is hoped that some University holiday flats will be available so that
participants can combine the meeting with a family holiday on the edge of the
Lake District and the Yorkshire Dales National Parks.

Further details and application forms are available from Dr Will Light,
Department of Mathematics, Lancaster University, Fylde College,
Lancaster LA1 4YF, England.

E-mail maa007@cent1.lancs.ac.uk


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

From: Joseph Wai Hung So <USERJOSO%UALTAMTS.BITNET@Forsythe.Stanford.EDU>
Date: Wed, 24 Jan 90 13:17:04 MST
Subject: Cluster Analysis Package

Does anyone know of any Cluster Analysis packages
for use on the MAC II ?

Joseph So
(Mathematics, U of Alberta. JOSO@UALTAMTS.bitnet).


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

From: Frank Stetzer <stetzer@csd4.csd.uwm.edu>
Date: 24 Jan 90 22:36:20 GMT
Subject: Lagrangian Dual Algorithm

I was rummaging thru my (paper) files and I found an old but (to me)
fascinating article from my grad school days:

Donald Erlenkotter, "A Dual-Based Procedure for Uncapacitated Facility
Location." _Operations Research_ V26 N6 (Nov-Dec 1978), pp992-1009.

I have lost track of this line of research, which I believe we referred
to as Lagrangian Duals. Does anyone have any more recent references to
this technique, especially as applied to facility location problems?
Email (preferably) or post. Many thanks.

Frank Stetzer stetzer@csd4.csd.uwm.edu
Statistical Software Consultant
Computing Services Division
U. Wisconsin - Milwaukee


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

From: Susan Mehringer <SIM%CORNELLF.BITNET@Forsythe.Stanford.EDU>
Date: Wed, 24 Jan 90 17:49:16 EST
Subject: Cornell Supercomputing Program for Undergraduate Research

Cornell National Supercomputer Facility
Supercomputing Program for Undergraduate Research

The Cornell National Supercomputer Facility (CNSF) at Cornell University
in Ithaca, New York is pleased to announce the third year of its
Supercomputing Program for Undergraduate Research. This program,
designed to prepare the computational leaders of tomorrow, provides
talented undergraduate students with an opportunity to develop
supercomputing skills while pursuing their own research using advanced
computational tools. The students' faculty advisers are also given
training and support as part of this program. Students and faculty
receive hands-on instruction in the use of the CNSF supercomputer and
the chance to interact with scientists in their fields of interest and
to participate in the research atmosphere of a national supercomputing
center.

The Program

Students and their advisers will participate in this program as teams,
with one faculty adviser and one or two undergraduate students per team.
Each team should propose a computational research project in any area of
science, mathematics or engineering. The project should be based on the
FORTRAN language, require the computational power of a supercomputer,
and be of a scope suitable for the student(s) to complete in the
following academic year.

The program begins with a four-week short course for undergraduates,
June 4 through 28, 1990, at Cornell. The students will be trained in
advanced computing methods, with particular emphasis on parallelism and
vectorization. Last year's program included numerical methods lectures
in areas such as Cholesky Factorization, Finite Difference Methods, and
Hyperbolic Equation Methods, and Science Applications, such as
"Circulation in Ancient Continental Oceans" and "Convection with
Magnetic Fields." Faculty advisers will join the students for the final
two weeks of the short course, June 18 through 28. The faculty members
will learn about new techniques and developments in supercomputing and
will participate in sessions on using advanced technology in the
undergraduate curriculum. Advisers and students will work together to
refine the computational strategy, advance the research, and further
develop the overall project plan.

The program continues after participants return to their home
institutions for the 1990-91 academic year. Students will carry out
their research projects by May, 1991, under the direction of their
faculty advisers. Each participant will be allocated ten service units
on the IBM 3090-600E for the computational research. The CNSF will
provide technical support for the duration of the project. Midterm and
final reports will be required.

Eligibility and Selection

A review committee composed of scientists from the Cornell faculty and
the CNSF staff will select approximately 15 teams to participate in the
program. Criteria for selection will include the merit of the
computational research and the academic credentials of the team members.
Each team's home institution must grant academic credit to participating
students, and provide network access to the CNSF or funding for dial-up
access. There is no limit to the number of teams that may apply from
one institution. Typical students will have completed the junior year
of study, but any qualified student with a graduation date no earlier
than December 1990 will be considered. Students must have completed
coursework in linear algebra, and coursework or experience in FORTRAN.
Applicants must be U.S. citizens or permanent residents. Special
consideration will be given to applicants from four-year colleges with
limited research facilities. Women, minorities, and disabled persons
are strongly encouraged to apply.

The Facilities

Participants will have access to all CNSF resources, including two IBM
3090 multiprocessor supercomputers with 12 vector facilities, parallel
processing, and extensive graphics facilities including specialized
animation equipment. The operating system, VM/XA with CMS, allows each
application to use up to a gigabyte of virtual memory. The FORTRAN
compilers support both vector and parallel processing. Most CNSF
facilities, including some of the graphics capabilities, are accessible
through dial-ups and networks.

Stipend and Logistical Details

Students who are accepted and complete projects will receive a stipend
of up to $2,000 plus a travel allowance. Faculty advisers will receive
a stipend of $250 /week of institute participation. The program
includes room and board for both students and faculty advisers in
Cornell University dormitories. Students and faculty will have access
to libraries and other resources on the Cornell campus.

This program is sponsored by the National Science Foundation and is
dependent on approval of funding through the Research Experiences for
Undergraduates Program of that agency.

How To Apply

Applications must be received by March 23, 1990. To obtain an
application form and more information, contact:

Donna Smith
Conference Coordinator
Cornell National Supercomputer Facility
Campus Road and Central Avenue
Cornell University
Ithaca, NY 14853-8301

Telephone: (607) 255-3985 or (800) 346-2673
Electronic mail to donna@tcgould.tn.cornell.edu.

Applicants should receive notification of the final selection by
April 9, 1990.


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

From: Melvyn Ciment <mciment@note2.nsf.gov>
Date: Thu, 25 Jan 90 11:41:17 -0500
Subject: Summer Institute in Japan

New SUMMER INSTITUTE in JAPAN for U.S.
Graduate Students in SCIENCE and ENGINEERING

The National Science Foundation (NSF) announces an
opportunity for 25 select U.S. graduate students in science
and engineering to spend the summer at Japanese government
research laboratories.

The goals of the new Summer Institute in Japan program are
to increase Americans' acquaintance with Japanese research
institutions, to promote the benefits of cooperative research
with Japanese investigators, and to build relations between
the U.S. and Japanese research communities. Those eligible
are graduate students (U.S. citizens or nationals) who are
enrolled at a U.S. institution in either a Ph.D. program
(persons in M.D. programs are not eligible) or an engineering
graduate program of which they have completed a year or more.

The 25 graduate students selected for this internship will
spend the summer participating in on-going research projects
at Japanese government laboratories in Tsukuba Science City,
about 60 km northeast of Tokyo. The interns will also
receive intensive Japanese language training and, through
lectures, discussions, and occasional field trips, compare
U.S. and Japanese culture, science policy, and the current
status of selected fields of science and engineering. Graduate
students travelling to Japan as interns in the Summer
Institute program will be in Japan for eight weeks from June
16 through August 7.

In all, 36 Japanese government research institutions are
potential hosts for the Summer in Japan interns. They
include such well known laboratories as the Electrotechnical
Laboratory, National Institute for Environmental Studies,
Mechanical Engineering Laboratory, National Institute of
Agrobiological Resources, University of Tsukuba, the National
Laboratory of High-Energy Physics (KEK), the National
Institute for Research in Inorganic Materials, and the
Tsukuba Life Science Center of the Institute of Physical and
Chemical Research (RIKEN).

The application deadline for the Summer Institute in Japan
program is March 1, 1990. NSF welcomes applications from all
qualified science and engineering graduate students, and
strongly encourages women, minorities, and persons with
disabilities to apply. Application materials may be obtained
from the Dean of the Graduate School on many campuses or from
NSF's Japan Programs Office, by telephone (202) 357-9558, or
through electronic mail to:

NSFJinfo@NSF (BitNet) or
NSFJinfo@note.nsf.gov (InterNet)


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

From: Nick Higham <MBBGSNH@CMS.MANCHESTER-COMPUTING-CENTRE.AC.UK>
Date: Thu, 25 Jan 90 14:27:34 GMT
Subject: Beyer Chair at University of Manchester

UNIVERSITY OF MANCHESTER
BEYER CHAIR OF APPLIED MATHEMATICS

Applications are invited for the Beyer Chair of Applied Mathematics
vacated by the retirement of Professor F. Ursell FRS. This is a
prestigious named Chair of Applied Mathematics and the University is
seeking to appoint an Applied Mathematician of distinction who will have
the opportunity to develop new research interests in the department,
while providing consolidation and leadership for the existing research
programme. A supporting lectureship is available under the New Academic
Appointments Scheme. Existing research is principally in the general
area of continuum mechanics, but applications are invited from
candidates of distinction in any area of Applied Mathematics.

Applications (one copy, suitable for photocopying) giving full details
of qualifications, experience, etc. and the names and addresses of three
referees should be sent to the Registrar, the University, Manchester M13 9PL,
UK, by February 23rd, 1990. Particulars of the appointment may be obtained
from the Registrar quoting reference 9/90. (Tel: 061-275 2028).
The University is an Equal Opportunity Employer.

K.E. Kitchen
Registrar


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

From: Philip Keenan <keenan@gargoyle.uchicago.edu>
Date: Fri, 26 Jan 90 09:09:59 CST
Subject: Vector Version of Paranoia

I am trying to diagnose floating point accuracy problems we are having
in our new vectorizing machine. I have run "paranoia", which seems to
be the standard test program for verifying IEEE standard behaviour in
scalar floating point machines. However, my copy of paranoia does not
vectorize, and our problems occur only in vectorized codes.

Does anyone know of a vectorizing version of paranoia, or of some
other program along the same lines, designed to carefully check
floating point arithmetic on vector (or parallel) machines?

Any program would be helpful, but I would especially be interested in
knowing if there is a standard program, like paranoia is in the scalar
case, which is generally accepted as canonical.

Replies may be sent to:
keenan@tartarus.uchicago.edu

Thanks very much!
-- Phil Keenan


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

From: Walter Gander <gander@inf.ethz.ch>
Date: 27 Jan 90 11:38:27+0100
Subject: Translation of Schwarz Textbook Recommended

The textbook on Numerical Analysis of H. R. Schwarz (University of
Zuerich) has recently been translated from German to English (John
Wiley, 1989). I'd like to recommend it to the na-community.

It contains the chapters linear equations, linear programming,
interpolation, approximation of functions, nonlinear equations,
eigenvalue problems, method of least squares, numerical quadrature,
ordinary differential equations and pde.

The book is widely used in German speaking countries and is rooted in
the tradition of Stiefel and Rutishauser.

Walter Gander

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

From: George Cybenko <gc@sp12.csrd.uiuc.edu>
Date: Sat, 27 Jan 90 15:38:25 CST
Subject: Supercomputer Applications Newsletter

Recently, I agreed to co-edit (with Dennis Duke of Florida
State University) the quarterly newsletter of
the IEEE Technical Committee on Supercomputer Applications.
The newsletter is much like the SIGNUM newsletter in that
typical issues contain some mix of technical short notes,
discussions and meeting announcements.

The Technical Committee focuses on supercomputer algorithms
and performance related issues. Position papers, commentaries
and benchmarking results are appropriate and will be considered
for publication. The TC has about 600 members.

If you have any material suitable for publication in the Newsletter,
please let me know. I would prefer to handle submissions by email
whenever possible. The next issue, the first that I am editing still
has some space, so a prompt response may make it into the next issue.

George Cybenko
Center for Supercomputing Research and Development
University of Illinois at Urbana
Urbana, IL 61801 USA

Phone : (217) 244-4145
email: g-cybenko@uiuc.edu


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

From: Tom Manteuffel <tmanteuf@copper.Colorado.EDU>
Date: Thu, 25 Jan 90 22:13:22 mst
Subject: Registration Form for Copper Mountain Conference

I have had a great number of requests for registration forms
for the Copper Mountain Conference on Iterative Methods. I
have attached a form to this message. Please note the student
registration fee that was absent from the previous announcements.

Tom Manteuffel


Conference Registration Form
Copper Mountain Conference on Iterative Methods
April 1-5, 1990

Name______________________________________________________________

Affiliation_______________________________________________________

Street Address____________________________________________________

City,State,Zip,Country____________________________________________

Office Phone_____________________Home Phone_______________________


Advanced Registration Fee (Postmarked by March 1, 1990)
( ) SIAM Member $135.00
( ) Non-member $150.00
( ) Student $ 25.00

Registration Fee (Postmarked after March 1, 1990)
( ) SIAM Member $160.00
( ) Non-member $175.00
( ) Student $ 35.00

Please Indicate method of payment:
( ) Check or Money Order enclosed
( ) Purchase Order (#________________From____________________)
( ) Master Card or Visa (Card#_______________________________)
(Exp. Date___________________________)

Authorized Signature______________________________


Payment must be made in U.S. funds drawn on a U.S. bank payable to
THE UNIVERSITY OF COLORADO.

Send this form with payment to: Division of continuing Education;
University of Colorado at Denver;1200 Larimer Street, Campus Box
164; Denver, CO 80204, USA.

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

End of NA Digest