NA Digest Sunday, September 17, 1989 Volume 89 : Issue 36

Today's Editor: Cleve Moler

Today's Topics:


From: David Levine <>
Date: Mon, 11 Sep 89 12:33:32 CDT
Subject: Argonne Parallel Programming Class

The Advanced Computing Research Facility (ACRF) at Argonne National Laboratory
is offering an introductory course on parallel computing December 6-8, 1989.
The date of the course is intended to make it convenient for those attending
the SIAM Conference on Parallel Programming being held the following week in
Chicago (the course is NOT associated with the conference). The course is
limited to 24 people on a first come first serve basis.

Topics to be covered include:

1) Parallelizing compilers.
2) The Monitor and Schedule packages for portable parallel programming.
3) Programming the Butterfly 2.
4) Programming the AMT DAP.
5) Programming the Connection Machine-2.
6) Introduction to the LAPACK project.

A portion of the third day will be devoted to each attendee's particular
project. The format of the course is alternating lectures and hands-on
work with the parallel computers in the ACRF. Fortran will be emphasized
as the primary programming language. Knowledge of Fortran and Unix will
be assumed.

Parallel computers currently in the ACRF are:

4-processor Ardent Titan
8-processor Alliant FX/8
16-processor Intel iPSC-VX hypercube
20-processor Encore Multimax
24-processor Sequent Balance 21000
32-processor Intel iPSC-1 hypercube
32-processor Butterfly TC2000
1024-processor Active Memory Technology DAP
16384-processor Thinking Machines CM-2

Those interested in the class should contact:

Teri Huml
Mathematics and Computer Science Division
Argonne National Laboratory
Argonne, IL 60439-4844
(312) 972-7163

There will be a $25.00 registration fee per person for universities,
federal laboratories and government organizations and $100.00 for
commercial organizations.


From: Michael Singer <>
Date: Mon, 11 Sep 89 12:17:45 EDT
Subject: Two Computer Algebra Workshops

CADE 90 CAP 90
Preliminary announcement

During the week Sunday, May 6, to Saturday, May 12, 1990, a
group of persons with a common interest are organizing two
contiguous workshops entitled:

``Computer Algebra and Differential Equations (CADE-90)''
(Sunday morning --- Wednesday noon) and

``Computer Algebra and Parallelism (CAP-90)''
(Wednesday afternoon --- Saturday afternoon)

at the Mathematical Sciences Institute (MSI) at Cornell University.
The CADE-90 organizers are Evelyne Tournier at IMAG in Grenoble and
Michael Singer at N.C. State University. The CAP-90 organizers are
Jean Della Dora at IMAG in Grenoble, John Fitch at the University of
Bath, and Erich Kaltofen at Rensselaer Polytechnic Institute.
Both committees include Richard Zippel at Cornell University.
We are combining efforts so that persons with overlapping interests
can stay the entire week.

The aim of CADE-90 is to allow workers in computer algebra to learn
of recent theoretical developments concerning differential equations
as well as make the researchers in theoretical areas aware of
questions arising in the design of computer algebra systems.
Peter Olver (University of Minnesota) will give four introductory
lectures on ``Lie Group Methods in Differential Equations". In
addition there will be a few invited one-hour lectures on topics such
as perturbation theory, summation and integration in finite terms,
effective methods in the theory of D-modules, differential galois
theory, and resummation techniques. In addition there will be one or
two presentations of software projects that integrate computer
algebra and differential equations (such as DESIRE). We plan to have
at most two lectures each morning and each afternoon, so that there
will be plenty of time for informal interaction.

The aim of CAP-90 is to allow workers in computer algebra to learn
the recent theoretical developments concerning the design of parallel
algebraic algorithms as well as make researchers working on
theoretical issues in parallel algebraic computation aware of
problems that arise when mapping these algorithms onto existing
parallel computers. The topics include vectorized long integer
arithmetic, parallel expression evaluation, parallel linear and
non-linear system solving, parallel sparse interpolation, parallel
polynomial root finding, and others. Furthermore, experience with
implementing parallel algebraic algorithms on parallel computers will
be discussed. The workshop will consist of several lectures by the
keynote speaker (unconfirmed) and by one hour lectures of several
invited researchers. The format of the workshop is such that
ample time for discussion sessions and impromptu presentations of
ongoing research is reserved. It is also anticipated that access to
the local Cornell supercomputers will be provided.

Funds are available for partial support of young predoctoral and
postdoctoral visitors. Anyone wishing to attend either workshop or
to obtain general information is encouraged to contact MSI through:
Valery Styles, Mathematical Sciences Institute, Cornell University,
201 Caldwell Hall, Ithaca, N.Y. 14853-2602. For more information
about the scientific content of CADE-90 contact Michael F. Singer,
Department of Mathematics, Box 8205, N.C. State University, Raleigh,
N.C. 27695-8205 USA (e-mail: SINGER@NCSUMATH.BITNET or, or Evelyne Tournier, TIM3-IMAG,
BP 68, 38402 St. Martin d'Heres cedex, FRANCE
(e-mail: TOURNIER@FRGREN81.BITNET). For more information about the
scientific content of CAP-90 contact Jean Della Dora, TIM3-IMAG, BP 68,
38402 St. Martin d'Heres cedex, FRANCE (e-mail: DELLADO@FRCICG71.BITNET),
John P. Fitch, School of Mathematical Sciences, University of Bath,
Bath BA2 7AY, ENGLAND (e-mail:, or
Erich Kaltofen, Department of Computer Science, Rensselaer Polytechnic
Institute, Troy, New York 12180-3590 (e-mail:


From: David Watkins <watkins@wsumath.bitnet>
Date: Tue, 12 Sep 89 10:54 PDT
Subject: Manuscript on Fundamentals of Matrix Computations

For some time I have been working on a manuscript entitled FUNDAMENTALS
OF MATRIX COMPUTATIONS that is intended to be used as an undergraduate
text or as a relatively easy graduate text. The book has been through
class testing at several institutions and is nearly in its final form.
One more round of reviews is planned. If you have taught matrix
computations and are interested in reviewing this book, please contact
my editor:
Valerie Hunter
Mathematics Editor
John Wiley and Sons
605 Third Avenue
New York, NY 10158
phone 212-850-6244

David Watkins, Washington State University (watkins@wsumath.bitnet)


From: Henry Wolkowicz <>
Date: Tue, 12 Sep 89 15:29:18 EDT
Subject: E-mail Group for Graduate Officers

I am trying to start an E-mail group for graduate officers
(chairman of graduate committees) at various institutions. I have tried
to collect several E-mail addresses and will continue collecting more.
I will try and be concise and short in all messages so as not to
increase (too much) the amount of junk E-mail we all get already.
I hope that this service will be useful to all. Please send
any items to me. I will collect and edit them and send them out in
a group.
I think that it would be helpful to pass information among the
different institutions with graduate programs. For example,
announcements of dissertation awards and deadline dates could be
passed and then posted. Any new information about Chinese students
coming or not coming would be helpful. Questions about details
of courses at different institutions could be passed.
Please send me the E-mail address of the graduate officers
at your institution and I will try and contact them.

Henry Wolkowicz; Department of Combinatorics and Optimization;
Faculty of Mathematics; University of Waterloo;
Waterloo, Ontario, Canada N2L 3G1 (519-888-4597 office; 746-6592 FAX)
{hwolkowicz@water.bitnet; }
{; usersunn@ualtamts.bitnet }


From: Ashwin A Desai <m2c!wpi!>
Date: 12 Sep 89 19:23:28 GMT
Subject: Simplex Method for Non-linear Optimization

I need any information I can get about the Nelder-Mead
procedure (a simplex method for minimization). I would
like to know if there are any books which explain
this method in some detail and also any details about
a software package for this method(if available).

Thanks in advance,
Worcester Polytechnic Institute


From: Jerry Johnson <>
Date: Tue, 12 Sep 89 12:36 CDT
Subject: Examples of Applications of Elementary Mathematics

D G Wilson <wilson@MSR.EPM.ORNL.GOV> asked in last week's newsletter
for "real live examples of the use of relatively elementary mathematics
(calculus, ordinary differential equations, vectors, elementary linear
algebra, etc.) in science and engineering. They want to use these
examples in the classroom to motivate concepts being taught. They
would appreciate hearing about any examples that you would be willing
to share."

A great idea ... and we would also appreciate receiving the same.
Please send contributions to:
Prof. Jerry Johnson, Mathematics Department,
Okla. State University, Stillwater, OK 74078.

Having asked, I will now give.

A spherical mine filled with an explosive of some given specific gravity
and encased in a one inch thick iron shell should float just submerged under
the ocean surface. What diameter should it have? This requires the solution
of a cubic equation. It can be used in calculus as a vehicle for applying
Newton's method, or in precalculus as motivation for looking at cubic
equations in the first place.


Date: Wed, 13 Sep 89 19:19 CDT
Subject: Positions at University of Iowa

Positions Available
The University of Iowa

The Mathematics Department of the University of Iowa invites appli-
cations for the following positions.:

1. Three tenure-track appointments at the Assistant or beginning
Associate Professor level beginning in the 1990-91 academic year.
One of these is to be in numerical analysis and two are to be
filled by specialists in harmonic analysis, probability theory, or
topology of manifolds. Selection will be based on evidence of out-
standing research accomplishments or potential, and teaching abil-
ity. A Ph.D. or equivalent training is required.
2. One senior faculty position beginning in 1990-91 academic year
or later. Only applicants of extraordinary stature will be consid-
ered. A strong record of leadership in teaching and research in
one of the department's current or developing areas of strength is
3. Pending availability of funds, one or more visiting positions
for all or part of the 1990-91 academic year. Selection will be
based on research expertise and teaching ability. Preference will
be given to applicants whose scholarly activity is of particular
interest to members of the current faculty.

Women and minority candidates are especially urged to apply for any
of the above positions. The University of Iowa welcomes the
employment of highly qualified professional couples on its faculty
and staff, permits the appointment of faculty couples within the
same department, and permits the sharing of a single appointment by
a faculty couple. Applications will be received until January 31,
1990, or until the positions are filled.

To apply send a complete vita and have three or more letters of
recommendation sent to:
Professor W. A. Kirk, Chair
Department of Mathematics
University of Iowa
Iowa City, Iowa 52242.

The University of Iowa is an Equal Opportunity/Affirmative Action


From: Richard W Brankin <>
Date: Sun, 17 Sep 89 10:47 GMT
Subject: NAG Response About Apollo Implementation

We were sorry to hear of the difficulties in attempting to re-compile the
NAG Fortran Library under the Apollo SR10.1 system by Mr. Calamai at the
University of Waterloo [NA News Digest, vol. 89, number 34, Sep 3rd 1989]
As far as we are aware, we had no contact with him prior to his public
message about the technical difficulties that he experienced. We have
responded to Mr. Calamai and below is an extract of our reply. We would
be most grateful if customers contact us first (by phone, fax, e-mail
or in writing) - to allow us the opportunity to comment and, if possible,
to assist - before using a public network to broadcast reports of apparent
difficulty involving the NAG Library; even if the Library itself is not
at fault (which we believe to be so in this case). We would certainly not
want recipients of such a message to infer (wrongly) that NAG does not provide
any advice or assistance.

We have been carrying out tests using Apollo SR10 with the NAG Fortran
Library. The initial tests, using a pre-release compiler (10.0), indicated many
problems with the Fortran compiler. A replacement Fortran compiler (10.6) came
with SR10.1 accompanied by a list of known errors. We have reported a number of
further compiler problems with SR10.1, and are awaiting a response from Apollo.
In the meantime we have been able to work around all of the problems. An
implementation of the NAG Fortran Library for use under SR10 is currently being
certified and we will release it when fully tested.

We have also tried Apollo's tools to convert object libraries for use between
SR9.7 and SR10. These tools break down after less than 300 modules have been
processed; again, Apollo have been informed.

NAG Technical Desk


End of NA Digest