NA Digest Sunday, December 27, 1992 Volume 92 : Issue 49

Today's Editor:

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

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: Anthony Skjellum <tony@helios.llnl.gov>
Date: Mon, 21 Dec 92 10:15:32 PST
Subject: Change of Address for Anthony Skjellum

Effective January 7, 1993, I will be leaving Lawrence Livermore
National Laboratory, and begin working at Mississippi State University
(MSU), Department of Computer Science, as Professor of Computer
Science. I am also associated with the National Science Foundation
Research Center for Computational Field Simulation, at MSU.

Here is my new professional address information [send to my attention]:

Mississippi State University
Department of Computer Science
Drawer CS [for non FEDEX]
300 Butler Hall, Corner of Perry & Barr [for FEDEX]
Mississippi State, MS 39762-5623
(601) 325-2756
FAX: (601) 325-8997
tony@cs.msstate.edu

or
Mississippi State University
NSF Engineering Research Center for Computational Field Simulation
PO Box 6176
Mississippi State, MS 39762

My efforts with "The Multicomputer Toolbox," "Zipcode," and related
research on 1st/2nd generation scalable concurrent libraries for
massively parallel computers, will continue uninterrupted at MSU.


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

From: Nick Gould <directory@directory.rutherford.ac.uk>
Date: Wed, 23 Dec 1992 14:05:14 GMT
Subject: Temporary Change of address for Nick Gould

I shall be spending a Sabbatical year at CERFACS in the sunny south of
France throughout 1993. My address will be

CERFACS, 42 Avenue Gustave Coriolis,
31057 Toulouse Cedex, France, EC
telephone: (33) 61 19 31 31
email: gould@cerfacs.fr

Queries concerning LANCELOT should be sent to me there.

Seasons greetings to all na-netters,

Nick Gould


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

From: John Mathews <mathews@fullerton.edu>
Date: 21 Dec 1992 06:53:00 -0800 (PST)
Subject: MATLAB Materials for Undergraduate Numerical Analysis

MATLAB materials for the undergraduate numerical analysis text
"Numerical Methods: for Mathematics, Science and Engineering"
2nd Ed, Prentice-Hall, Inc. 1992, are available from P-H, Inc.
1. "MATLAB Programming Guidebook for Numerical Methods."
2. IBM-PC and Macintosh disks containing MATLAB programs
which implement all the algorithms in the text.

To obtain a copy these materials for your course call Prentice Hall
Customer Service for Professors: 1-800-526-0485.

My E-mail address is in the preface and I will supply
technical support the MATLAB materials for my text.

John Mathews
Prof. of Mathematics
Calif. State Univ. Fullerton
Fullerton, CA 92631
(714) 773-3631
e-mail mathews@fullerton.edu


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

From: Allison Bogardo <bogardo@siam.org>
Date: Mon, 21 Dec 92 20:12:39 EST
Subject: SIAM Student Paper Competition

SIAM Student Paper Competition

The annual SIAM Student Paper Competition will be held during the 1993 SIAM
Annual Meeting.

If you are a student or know of a student who should be included in the
competition, here are the details:

The authors of the three best papers in applied and computational mathematics
written by students and submitted to SIAM will be invited to attend the 1993
annual meeting in Philadelphia, July 12-16. Each winner must present his/her
paper at the meeting and will receive a $750 cash award as well as gratis
registration for the meeting. Winners will be awarded calligraphed
certificates at a special prize ceremony at the meeting. Papers must be
singly authored and not previously published or submitted for publication to
be eligible for consideration. To qualify, authors must be students in good
standing who have not received their PhDs at the time of submission.

In submitting their work for publication, authors are asked to consider SIAM
journals.

Submissions must be received by SIAM on or before March 15, 1993.

Submissions, which must be in English, can be sent by regular mail or fax.
Each submission must include (1) an extended abstract NOT LONGER THAN 5 PAGES
(including bibliography); (2) the complete paper, which will be used solely
for clarification of any questions; (3) a statement by the student's faculty
advisor that the paper has been prepared by the author indicated and that the
author is a student in good standing; (4) a letter by the student's faculty
advisor describing and evaluating the paper's contribution; and (5) a short
biography of the student.

Submissions will be judged on originality, significance, and quality of
exposition.

The winners will be notified by May 28, 1993.

If you have any questions, please contact:

Allison Bogardo
SIAM
3600 University City Science Center
Philadelphia, PA 19104-2688
Telephone: (215) 382-9800
E-mail to bogardo@siam.org
Fax to (215) 386-7999


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

From: D. F. Griffiths <dfg@mcs.dundee.ac.uk>
Date: Mon, 21 Dec 92 17:33:20 GMT
Subject: Dundee Conference 93

15th BIENNIAL CONFERENCE
ON
NUMERICAL ANALYSIS
UNIVERSITY OF DUNDEE
SCOTLAND, UK
Tuesday 29 June - Friday 2 July 1993

Address for correspondence

Dr D F Griffiths
Numerical Analysis Conference
Department of Mathematics and Computer Science
The University, Dundee, DD1 4HN
Scotland, UK

Telephone: 0382-23181 Ext 4467/4471
FAX : 0382-201604
email: na.griffiths@na-net.ornl.gov
OR : dfg@uk.ac.dund.mcs

The conference will feature a special invited lecture in honour of A R
Mitchell and will be presented by M J D Powell. The other Principal
Speakers will include

J W Barrett I S Duff C M Elliott
P Gill D J Higham N K Nichols
P Townsend J M Sanz-Serna M N Spijker
G W Stewart A M Stuart R Temam
M J Todd

Brief Communications
A limited number of short (20 minute) papers will be
accepted for presentation. Prospective contributors
should submit a title and short abstract as soon as
possible and no later than 21 May 1993. Please indi-
cate on the registration form whether or not you wish
to present a paper.

Registration Fee
The registration fee will be #50.00 (# = Pounds
Sterling). This includes the cost of morning coffee and
afternoon tea.

Proceedings
The proceedings of the invited talks only will be pub-
lished. Copies may be ordered on payment of #16.00 at
the time of registration. Later requests for copies
should be made to the publisher.

Accommodation
Accommodation for the conference will be provided in
West Park Hall, a student residence which consists
mainly of single study bedrooms. Full board (bed,
breakfast, lunch and dinner) is available from Monday
night to Friday evening at a cost of #136.00. Please
note that this does not include dinner on Thursday
night, when the conference dinner will be held.

In addition, bed and breakfast will be available on the
nights of Sunday 20 June and Friday 2 July (#19.00)
as will dinner on Monday night (#8.00).

Information on local hotels can be provided on request.

Conference Dinner
The conference dinner will be held on the Thursday
evening in West Park Hall. The cost, inclusive of wine
will be #15.00. The number of places available at the
conference dinner is limited and will be allocated
according to the date of booking.

To obtain an electronic registration form, please contact
na.griffiths@na-net.ornl.gov
OR
dfg@uk.ac.dund.mcs
The completed registration should arrive no later than 28 May 1993.

A R Mitchell Fund

The 1991 Dundee Numerical Analysis Conference celebrated the 70th birthday
of A R Mitchell. In addition to his distinguished contributions to
numerical analysis extending over many years, Ron has been known for the
outstanding encouragement he has given to young numerical analysts and the
Organizing Committee therefore felt it appropriate to mark his birthday by
instituting a special fund to provide assistance to young numerical analysts
to attend Dundee meetings.

There are no formal means of applying for support from the fund but the
organisers welcome enquiries from anyone who feels that they may be
eligible. Allocation of funds will be made on the basis of need, and also
on the basis of abstracts of papers intended for presentation at the
conference.

We would like to thank those who have contributed generously to this fund
and we again appeal to colleagues for contributions. Cheques should be made
payable to "The University of Dundee: A R Mitchell Fund".


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

From: Simon Chandler-Wilde <s1186@bradford.ac.uk>
Date: Wed, 23 Dec 92 21:11:50 GMT
Subject: Northern Universities' Numerical Analysis Day

NORTHERN UNIVERSITIES' NUMERICAL ANALYSIS DAY
Thursday 7th January
University of Bradford

Outline of the Programme

Meeting begins at 10 a.m.

"An Optimal Positive Definite Update for Sparse Hessian Matrices"
Roger Fletcher, University of Dundee

"Acceleration of Convergence of a Sequence of Vectors"
Peter Graves-Morris, University of Bradford

"A Pseudospectral 3D Boundary Integral Method Applied to a
Nonlinear Model Problem From Finite Elasticity",
Ivan Graham, University of Bath

"Singularities of Solutions of Abel-Volterra and Abel-Fredholm
Integral Equations", Donald Kershaw, University of Lancaster

"Boundary Integral Methods for the Helmholtz Equation in
Exterior Domains", Sia Amini, University of Salford

"An Iterative Variant of a Chebyshev Collocation Method for
Symm's Integral Equation on an Open Arc"
Jeremy Levesley, University of Leicester

For further details contact Simon Chandler-Wilde,
Department of Civil Engineering, University of Bradford
(Email: S.N.Chandler-Wilde@uk.ac.bradford)


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

From: George Cybenko <gvc@cosmo.dartmouth.edu>
Date: Tue, 22 Dec 1992 21:11:16 -0500
Subject: Symposium on Parallel I/O and Databases

Dartmouth Institute for Advanced Graduate Studies, 1993

Issues and Obstacles in the Practical Implementation of
Parallel Algorithms and the Use of Parallel Machines

Symposium: Parallel I/O and Databases

School in Parallel Programming

June 21-June 30, 1993 Hanover, New Hampshire


This summer, Dartmouth will hold its second DAGS institute to
promote the use of high performance computing. As in 1992, when
it was established with the support of the NSF and Dartmouth Col-
lege, the institute will bring together researchers and practi-
tioners from academia and industry. This year, the institute
will be composed of two parts: a symposium (June 21-23) of invit-
ed and contributed talks, followed by a hands-on school (June
24-30) on parallel programming.
Housing will be available at economical rates in college dor-
mitories, near to conference sessions and centrally located on
the Dartmouth campus.

June 21-23: SYMPOSIUM: The focus of the symposium this year
will be on parallel I/O and databases . Invited speakers are:
Alok Aggarwal (MIT), Garth Gibson (Carnegie Mellon University),
David Scott (Intel Supercomputers), Jeffrey Vitter (Duke), David
Waltz (Thinking Machines), John Wilkes (HP).

A limited number of contributed papers will be selected for
presentation. Although the focus of the invited speakers is on
parallel I/O and databases, we welcome papers from any of the fol-
lowing topics:
parallel I/O systems and algorithms,
parallel database systems,
software issues in the construction of parallel programs,
parallel programming environments and tools,
implementation issues of parallel algorithms,
visualization algorithms and tools.

Authors are invited to send 10 copies of their doubled-spaced
manuscripts to be received by February 15, 1993. Manuscripts
should be limited to 20 pages. Contributors will be notified of
acceptance by March 15, 1993 and will be given instructions for
the final presentation of their paper at that time. Please send
research papers to Debra Minichiello, 6188 Bradley Hall, Dart-
mouth College, Hanover, NH 03755-3551 (email:
debra.minichiello@dartmouth.edu).

June 24-30: SCHOOL ON PARALLEL PROGRAMMING: The school will
involve a hands-on course on programming parallel algorithms us-
ing the NESL parallel language. The course will be
taught by the developer of NESL, Professor Guy Blelloch of
Carnegie Mellon University. The course will introduce several
parallel data structures and a variety of parallel algorithms and
then look at how they can be implemented.
The school will include other activities conducted by different
faculty of the DAGS '93 school, such as tutorials on various
software issues of parallel programming and exposure to or de-
monstration of other parallel programming languages, such as C*.

Program Committee: Guy Blelloch, Tom Cormen, George Cybenko,
Phil Hatcher, Donald B. Johnson, David Kotz, Fillia Makedon
(chair), Panagiotis Metaxas, Grammati Pantziou, Eric Schwabe,
Clifford Stein, Clifford Walinsky.

Organizing Committee: Fillia Makedon (chair), Donald John-
son, Ken Bogart, Tom Cormen, Wayne Cripps, George Cybenko, Scot
Drysdale, Mark Franklin, David Kotz, Don Kreider, Panagiotis
Metaxas, Debra Minichiello, Grammati Pantziou, Cliff Stein, Clif-
ford Walinsky, Patricia Wilson.

For more information please contact:
Prof. Fillia Makedon, Dept of Mathematics and Computer
Science, Dartmouth College, Hanover, NH 03755-3551 (e-mail:
makedon@dartmouth.edu). To be put on our mailing list, please
email dags@dartmouth.edu.


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

From: I. H. Sloan <sloan@hydra.maths.unsw.EDU.AU>
Date: Tue, 22 Dec 92 13:00:01 +1100
Subject: Position at New South Wales

UNIVERSITY OF NEW SOUTH WALES
SCHOOL OF MATHEMATICS
RESEARCH ASSOCIATE

Applications are sought for a Research Associate position supported by
an Australian Research Council program grant held by Professor I H
Sloan. The program is concerned with the development and justification
of numerical methods for boundary integral equations, multiple integration
and other aspects of boundary value problems.

Applicants should have a PhD (or equivalent) in an appropriate discipline,
and proven research accomplishment. Knowledge of one or more of
integral equations, numerical analysis, partial differential equations,
analysis, algebra and computing would be an advantage.

The position is available initially for 1 year, with a prospect of renewal.

Applicants should include the names of three academic referees with their
application.

Applications close: February 10, 1993

Salary is in the range: $34,953 to $38,950 per annum.

Further information from Dr R S Womersley on (02) 6972998 (office),
(02)6626445 (fax), R.Womersley@unsw.edu.au (email).


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

From: M. Baines <M.Baines@reading.ac.uk>
Date: Tue, 22 Dec 92 15:45:21 GMT
Subject: Chair of Applied Mathematics, University of Reading, UK

Applications are invited for the Professorship of Applied Mathematics
in the Department of Mathematics at the University of Reading, UK.
Applications from those whose interests lie in the broad area of
fluid mechanics (including computational fluid dynamics) will be
particularly welcome.

Closing date 18th January. For further information telephone

44 734 318115
(UK)


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

From: James Demmel <demmel@zil.CS.Berkeley.EDU>
Date: Tue, 22 Dec 92 22:41:05 -0800
Subject: Position at Berkeley in Scientific Computing

Two Year Position available
as a
Visiting Postdoctoral Research Engineer
at the
Computer Science Division
University of California at Berkeley

An HPCC/NASA-funded postdoctoral position in scientific computing is
available to support work on parallelizing an Earth Systems Model, a
model of the earth's climate, including coupled models of atmospheric
dynamics, atmospheric physics, atmospheric chemistry and ocean dynamics.
It is the only model of its kind, and will be used for studies of ozone
depletion and other global climate changes over future decades. This is
a collaborative effort among atmospheric, oceanic, geophysical and computer
scientists at

Colorado State (Randall)
JPL (Halpern)
Lawrence Livermore (Wuebbles, Dannevik)
Princeton (Philander)
UC Berkeley (Demmel, Stonebraker),
UCLA (Arakawa, Mechoso, Neelin, Turco)
UCSB (Dozier)

At Berkeley, Prof. Demmel is responsible for parallelizing the atmospheric
chemistry model developed by Turco at UCLA, and integrating the four numerical
models; this is the activity for which the position is available.
Prof. Stonebraker is responsible for the database (this activity is also
known as the Sequoia project). At Berkeley we have our own CM-5, and have
remote access to whatever parallel processors we will need.

This position will largely involve implementation, testing and measurement of
large codes on a variety of parallel machines, but initially the CM-5. The
ideal candidate would have a recent PhD in Computer Science or Numerical
Analysis, have significant parallel programming experience in scientific
computing, have a background in numerical methods, and enjoy interacting
with scientists in other disciplines. The candidate should have a committment
to building working software on computers not known for their friendly
programming environments in order to help understand one of today's most
important scientific problems: understanding global climate change.
Applications are invited from well-qualified candidates who might not
formally meet all these requirement.

This appointment is for two years, with a possible one-year extension to be
determined later, starting as soon as possible after February 1993.
Applications should be sent to

Prof. James Demmel
Computer Science Division
University of California
Berkeley, CA 94720
(510)643-5386
demmel@cs.berkeley.edu

The University of California is an affirmative action/equal opportunity
employer.


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

From: George Cybenko <gvc@cosmo.dartmouth.edu>
Date: Tue, 22 Dec 1992 21:11:16 -0500
Subject: Positions at Dartmouth

Dartmouth College
Thayer School of Engineering

Faculty Positions in Computer Engineering

The Thayer School of Engineering invites applications for
computer engineering faculty positions at the assistant
or associate professor level. Requirements include an earned
doctorate in computer engineering, computer science,
electrical engineering or related field. In addition,
applicants should present evidence of excellence, or potential
for excellence, in teaching and research. All areas of computer
engineering will be considered but specialization in high
performance computer systems, scientific computing, computer
communications or graphics/visualization technology will be
given preference.

Responsibilities include teaching at the undergraduate and graduate
levels, conducting funded research and advising graduate students
in the Engineering and Computer Science Ph.D. programs.
Compensation is competitive and commensurate with qualifications.

The Thayer School, founded in 1867, is the oldest
professional school of engineering in the United States. Together
with the Amos Tuck School of Business
Administration and the Dartmouth Medical School,
Thayer is one of Dartmouth's three professional postgraduate schools.
Faculty interests at Dartmouth currently include visualization for scientific
and medical computing, performance analysis, scientific computing,
rapid prototyping, and parallel algorithms.
The campus is located in a small New England town with
airline service to Boston and New York, interstate highway
access (two hours by car from Boston, three hours from Montreal)
and Amtrak rail service to New York and Montreal. The area is
known for its excellent outdoor and cultural activities.

Applications with current resumes and the names of
at least four references should be sent to:

Professor George Cybenko
Computer Engineering Search Chair
Thayer School of Engineering
Dartmouth College
Hanover, NH 03755-8000 USA.


Applications received prior to February 29, 1993 will be given
full consideration. Dartmouth College is an Equal
Opportunity /Affirmative Action employer and encourages
applications from women and members of minority groups.


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

From: Todd Torgersen <torgerse@mthcsc.wfu.edu>
Date: Mon, 21 Dec 92 14:23:48 EST
Subject: Report on the Twelvth Parallel Circus

The Twelvth Parallel Circus

The Twelfth Parallel Circus was held on October 30-31 at the North
Carolina Supercomputer Center. The Parallel Circus is an informal
meeting, with the program planned on the morning of the first day.
Many long-time attendees of the Circus commented that this was the
``best Circus ever'' in terms of organization and quality of the
presentations. A most enjoyable time was had by all. Many thanks
to the organizers, Tim Kelley, Bruce Loftis, Bob Plemmons and Gene
Golub, with special thanks to moderator Gene Golub and to local
host Bruce Loftis at the supercomputing center.

The talks were primarily concerned with ``parallel numerical algorithms''
and ``parallel software and hardware performance.'' Some speakers have
provided abstracts of their presentations and these are given in an appendix.
Many thanks to Chuck Baldwin, whose preliminary report appears (with some
minor editing) in the section below.

Summary of Presentations
------------------------

Friday started out with a number of talks related to eigenvalues.
The first, entitled ``Parallel Eigensolvers Based on Homotopy Continuation''
was given by Zhonggang Zeng from Northern Illinois. This talk focused on
using ODE methods to compute eigenvalues of an unknown matrix from those of
a closely related matrix whose eigenvalues are either known or easily
computed.
The second talk of the day was by Charles Fulton of the Florida Institute
of Technology and was titled ``Parallel Computation of Spectral Density
Functions for Sturm-Liouville Problems.'' This talk described a package
called SLEDGE - Sturm-Liouville Estimates Determined by Global Errors --
and it's implementation on an Intel iPSC/860.
Gary Howell, also of the FIT, next discussed ``An Adapted Bandwidth
Algorithm for Stable* Reduction to (Near) Tridiagonal Form.'' His talk
focused on using Gaussian Transforms to obtain solutions to Sylvesters
Equation, AX + XB = C.
Fourth up was a talk by Kevin Gates from ETH on ``A Rank-Two Divide
and Conquer Algorithm for the Symmetric Tridiagonal Eigenvalue Problem.''
He presented the basics of an algorithm similar to Cuppen's algorithm,
but is based on rank two updates. Numerical comparisons of several
algorithms were given and show the new algorithm to be competitive.
``A Homotopy Method for Solving Riccati Equations on a Shared Memory
Parallel Computer'' by Dragan Zlgic of Virginia Tech was next up. This
work was also concerned with using Homotopy methods for finding eigenvalues
of a matrix by using a related matrix and solving the Riccati equations,
AX + XA^T - XX + R = 0.
Christopher Sprenger presented a two part talk entitled ``Parallel
QR/Sciddle: A System for Asynchronous Remote Procedure Call''.
Gong Chen was next with a talk titled ``A Highly Parallel Algorithm
for Hessenberg Reduction.'' This talk described work done on an Intel
iPSC/860 to find methods for reducing matrices to Hessenburg form by
partitioning schemes suited to distributed memory type machines and
follows work done by Dongarra and van de Geijn.
George Pitts next gave a talk titled ``Strategies for Parallelizing
PDE software.'' This work investigated using Explicit Parallelization,
Automatic Parallelization, and a mixture of the two. Performance-critical
``kernel'' routines are hand coded and applied to various routines from
ELLPACK.
``Parallel Direct Solution of Indefinite Systems of PDEs Discretized with
Finite Element/Spectral Methods'' by Reza Mehrabi of MIT was next up. Here
a crystal growth program was ported to the iPSC/860 which used direct methods
along with nested bisection in 2D to obtain Finite Element solutions.
``A New Family of Preconditioners for Domain Decomposition'' by Mo Mu
of Purdue was next up. This work involved using rational functions as
preconditioners for elliptic problems with two different regions in 2D.
Victor Eijkhout of Tennessee was next with ``Observations about
Distributed Iterative Methods.'' This work described work done on
parallelizing iterative solvers on distributed memory machines and gives
a new variant of CG which reduces synchronization effects.
``Parallel Evaluation of Rational Matrix Functions'' was presented
by Lothar Reichel of Kent State. This presentation was on joint work
with Stratis and D. Calvetti. The work utilizes Partial Fraction
Representations to solve problems derived from elliptic operators and
considers some problems and possible solutions which involve poles which
are close in the partial fraction representation.
The last talk of the day was ``PYRROS: Static Scheduling and Code
Generation for Parallel Programs'' by the ETH group in Switzerland.
This was an overview of the PYRROS project and included discussions of
how PYRROS maps tasks to processors and schedules those tasks.

During the day (Friday) we had a tour of the North Carolina Supercomputer
Center in which we were introduced to the new 32 processor Kendall Square
machine at NCSC. The new machine utilizes a ring topology and, with
multiple rings scales to 1088 processors. The machine is up and running,
but some refinement to the software environment is in progress.

Saturday was a short day and involved several software related topics.
The morning started out with ``Prototyping Parallel Algorithms'' by Lars
Nylend of UNC Chapel Hill. This talk detailed a system called Proteus
which is a language for parallel programming with specific attention paid
to data parallelism, and loosely synchronous parallel execution - an
example program of the N body problem was given.
``Scheduling Regular and Irregular Computations on the CM-5'' by Rajeev
Thakur was up next. This describes an algorithms for global communications
on the CM-5 with superior performance.
Next on the schedule was ``A Proximal Based Decomposition Method
for Convex Minimization'' by Gong Chen of Maryland. This involved using the
fact that convex minimization was equivalent to solving a mini-max problem
via a new method that the authors are developing.
``Component Analysis Based on Serial Results for Analyzing Parallel
Iterative Performances'' by Shu-Mei Richman of the University of Virginia
was next. This was an investigation of GMRES and BiCG based on parameters
usually involved in iterative methods (Preconditioning methods, residual
tolerance, problem size, number of processors).
``Distributing Symbolic Computations'' was next given by Todd Torgersen
of Wake Forest. An overview of ALTS - a system for distributing symbolic
computations was given.
``Improvement of Communications Bandwidth Using the Forced Message Type''
by Xinglai Zhuang of Mississippi State was next. This involved a study of
various global communication operations using two different methodologies
on the Intel iPSC/860.
The program continues with ``Solving Large Systems of Linear Equations
using Asynchronous Iterations on Distributed MIMD Computers'' by J. Nieplocha
of the University of Alabama. This was a overview of issues which regularly
come up in asynchronous communications on MIMD computers such as communication
load, convergence of an algorithm, and load balancing.
Robert White of NC State next gave a talk on general issues related to
parallelizing grand challenge codes, the convection-diffusion type equation
was used as an example.
``A Parallel Algorithm to Calculate Smooth Boundary Conditions for a Domain
Decomposition Method'' by Kelly Black of NC State was up next. This talk
examined using Chebyschev Collocation on parabolic problems with interfaces
on the interior of a region.
``Orbit Prediction on Parallel Computers'' by Beny Neta of the Naval
Postgraduate School was next on the hit parade. This was a talk of using
an Intel iPSC to develop more parallel algorithms for accurately predict
orbits.
Last was Roldan Pozo of the University of Tennessee with ``Object
Oriented Scalable Linear Algebra in C++''. This was a talk of work done
in connection with the LAPACK project to have an object oriented
version available soon and how it could be of use to people.

=============================================================================
|| The Thirteenth Parallel Circus is planned for April at the University ||
|| of Illinois CSRD in Champaign-Urbana. ||
=============================================================================

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

From: SIAM <ddilisi@siam.org>
Date: Wed, 23 Dec 92 09:37:39 EST
Subject: Season's Greetings from SIAM

'
SSS EEEEE A SSS OOO N N SSS
S S E A A S S O O N N N S S
SS EEE A A SS O O N N N SS
SS E AAAAA SS O O N N N SS
S S E A A S S O O N NN S S
SSS EEEEE A A SSS OOO N N SSS


GGGGGG RRRRR EEEEE EEEEE TTTTTTT IIIII N N GGGGGG SSS
G R R E E T I N N N G S S
G R RRR EEEE EEEE T I N N N G SS
G GG R R E E T I N N N G GG SS
G G R R E E T I N NN G G S S
GGGGGG R R EEEEE EEEEE T IIIII N N GGGGGG SSS


FFFFFF RRRRR OOO MM MM
F R R O O M M M M
FFFF R RRR O O M M M
F R R O O M M
F R R O O M M
F R R OOO M M


SSS IIIII A MM MM
S S I A A M M M M
SS I A A M M M
SS I AAAAA M M
S S I A A M M
SSS IIIII A A M M




[ ... and from all of us at NA Net. -- Cleve]


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

End of NA Digest

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