NA Digest Sunday, September 3, 1989 Volume 89 : Issue 34

Today's Editor: Cleve Moler

Today's Topics:


From: Melvyn Ciment <>
Date: Mon, 28 Aug 89 14:08:57 -0400
Subject: NSF Division of Advanced Scientific Computing

National Science Foundation
Division of Advanced Scientific Computing

The goal of the New Technologies Program is to encourage
fundamental research related to the use of high-performance
computing in solving scientific and engineering problems.


High-performance computing refers to the full range of
supercomputing activities including existing supercomputer
systems, special purpose and experimental systems, and the new
generation of large scale parallel architectures. Topic areas
include software development for high performance computing
environments, performance evaluation, comparison of systems and
architectures, graphics, and visualization.

Scientific and Engineering Applications

The program supports research leading toward prototype
implementations of algorithms, models and environments for
end-user applications. Areas include, but are not limited to,
linear algebra, solution of partial differential equations, and
non-linear phenomena at a scale where supercomputers or massively
parallel systems are required. Application areas are drawn from
all the disciplinary areas supported by NSF.

Innovative Support Systems

High-performance computing systems frequently entail support
requirements unique to their environments. Among them are
requirements for output media, such as graphics or downloading to
work stations, for peripheral mass-storage which is organized
specifically for effective use by scientists and engineers, and
for convenient and effective interactive use. The program
encourages architectural and software developments in these

Scientific and Engineering Software for High-performance Systems

New languages, compilers and operating systems for scientific use
of supercomputers and massively parallel systems are appropriate
areas for support. Additionakky, means of improving performance
or functionality of existing software systems provide a basis for
extending their utility. Related redearch in performance
evaluation and benchmarking is considered to fall within these


Projects suitable for funding under the New Technologies Program
will have well-defined research goals; those that identify new
research issues are especially encouraged. Novelty of approach
and development of new methodology should be stressed.
Theoretical, experimental, and/or computational research may be
proposed. Proposals for engineering studies ioncluding the
construction of prototype systems will also be considered. In all
cases, the relationship to high-performance computing should be
made explicit in the proposal.


Proposals are accepted throughout the year.


Telephone, mail and electronic mail inquiries about the New
Technologies Program are welcome. They should be addressed to

Dr. Nathaniel Macon
Program Director
New Technologies Program
Division of Advanced Scientific Computing
National Science Foundation
Washington, DC 20550
tel. 202-357-7727

E-mail; nmacon@note,


From: Bob Ward <ward@rcwsun.EPM.ORNL.GOV>
Date: Mon, 28 Aug 89 15:14:19 EDT
Subject: Householder Fellowship at ORNL


Mathematical Sciences Section
Oak Ridge National Laboratory

The Mathematical Sciences Section of Oak Ridge National Laboratory
invites outstanding candidates to apply for the 1990 Alston S.
Householder Fellowship in Scientific Computing.

Alston S. Householder was the organizer and founding Director of the
Mathematics Division (precursor of the current Mathematical Sciences
Section) at Oak Ridge National Laboratory (ORNL). In recognition of
the seminal research contributions of Dr. Householder to the fields of
numerical analysis and scientific computing, a distinguished postdoctoral
fellowship program has been established at ORNL and named in his honor.
The Householder Fellowship is supported by the Applied Mathematical
Sciences Subprogram of the U.S. Department of Energy.

The purposes of the Householder Fellowship are to promote innovative
research in scientific computing on advanced computer architectures and
to facilitate technology transfer from the laboratory research
environment to industry and academia through advanced training of new
computational scientists. The Householder Fellowship is for a term of
one year, renewable for a second year. Benefits of the Fellowship
include a competitive salary, fringe benefits, travel opportunities,
access to state-of-the-art computational facilities (including both
parallel architectures and high-performance personal workstations), and
collaborative research opportunities in a very active research program
in advanced scientific computing. Competition for the appointment is
open to U.S. citizens and permanent residents. Applicants should have
completed a doctoral degree in computer science, mathematics, or
statistics within three years prior to the appointment and have a
strong background and research interest in large-scale scientific

The Mathematical Sciences Section of ORNL has research programs in
Computational Mathematics, Computer Performance Characterization,
Applied Analysis, and Computational Statistics. The precise research
emphasis of the Householder Fellow would necessarily depend to a great
degree on the research interests of the selected Fellow. Areas of
particular interest at ORNL, and in which applicants would be
especially encouraged, include:

1. Computational linear algebra, with special emphasis on sparse matrix
computations on advanced computer architectures.

2. Partial differential equations, with special emphasis on the
development of novel algorithms for solving mathematical problems arising
in environmental cleanup, such as fluid flow through porous media.

3. Tools for the development and analysis of parallel programs,
including programming environments for parallel computers and methods
for measuring and modeling the behavior and performance of parallel

4. Computational statistics, with special emphasis on the development
of procedures and algorithms for use in the design and analysis of
computational experiments, computer-aided experimental design, and
large-scale statistical problems.

5. ``Grand Challenges'' in computational science, emphasizing the use
of advanced computer architectures to solve important problems in
science and engineering in which fundamental breakthroughs are
dependent on unusually large computational requirements that tax the
capacity of current supercomputers.

Applicants should send a resume, statement of research goals, and the
names of three references to James Atherton, PhD Employment, Oak Ridge
National Laboratory, P.O. Box 2008, Oak Ridge, TN, 37831-6216, marked
``Attn: Householder Fellowship.'' The deadline for applying is
December 1, 1989, and the selection committee's decision on the winning
candidate will be announced in January 1990. The position will
commence in 1990.

For further information contact Robert C. Ward by phone at 615-574-3125
or by electronic mail at


From: Bradley Dickinson <bradley@ivy.Princeton.EDU>
Date: Mon, 28 Aug 89 22:36:08 EDT
Subject: New Telephone Numbers for Princeton

Princeton University is getting a new phone system and the exchange for
university numbers will change from 452 to 258 - effective Sept. 1, 1989
at 5 pm EDT. This new exchange will also cover some numbers that have
previously been on the 987 exchange (I think this was used for recent
overflow.) People should update their phone directories accordingly.
For example, my number becomes (609)-258-4644 and the FAX number for
the School of Engineering and Applied Science becomes (609)-258-6744.

Brad Dickinson


Date: Mon, 28 Aug 89 15:54:25 EMT
Subject: "Last" Loen Conference

JUNE 3 -- 7, 1991
in Norway

Organised by:

Jack Dongarra (ANL, Argonne), Iain Duff (Harwell Laboratory),
Patrick Gaffney (BSC) and Sean McKee (U. of Strathclyde).


Environmental Sciences
Reservoir Simulation & Seismic Processing
Large Scale Scientific Computations
Pattern Recognition/Image Processing & Computer Vision
Parallel Programming Languages - such as ADA
Parallel Numerical Algorithms


organized by Mary Wheeler (U. of Houston/Rice University)


Social Events will be arranged
Student Scholarships will be available
There will be a special surprise for those conference
attendees who have attended all three ``Loen'' conferences


From: Bob Mattheij <>
Date: Tue, 29 Aug 89 16:22:32 -0200
Subject: Stiff ODE Solver in Pascal

Does anyone know of a good implementation of
a BDF based method for stiff ODE, written in

Bob Mattheij
Dept of Math
Eindhoven University of Technology
PO Box 513
5600 MB Eindhoven
The Netherlands



From: Alan McKenney <>
Date: Tue, 29 Aug 89 16:03:44 EDT
Subject: New Yorker Articles on Radiation Dangers

In NA Digest 89:29 ( Sunday, July 30, 1989 ), George Byrne
<GDBYRNE%ERENJ.BITNET@Forsythe.Stanford.EDU> writes:

> Subject: New Yorker Articles on Radiation Recommended
> The New Yorker Magazine had an article in three parts on electromagnetic
> radiation, which may be of some interest to NA-NET readers/users.
> The articles are not lightly written; but they may be important to us for
> health reasons.
> George Byrne

I read these articles (or part of them, anyway) when they came
out. *If true*, there would be cause for concern. I put the *If* in
there, because I don't have faith in the _New_Yorker_'s zeal for
accuracy. They have in the last few years printed two articles that I
know of (there are probably more that I don't know of) which many
reputable and knowledgeable people disagreed with, both as to the facts
and to the interpretation. These disagreements I found out about from
other sources, since the _New_Yorker_ has no letters column and, if it
has ever printed any rebuttal or anything which disagreed with a
previous article, I have not seen it in 3--4 years of reading. This
suggests to me an editorial policy of printing "interesting reads", as
opposed to participation in the process of sifting out truth from

In fact, I have the impression that this series *has* been
disputed. The VDT study, in particular, I seem to remember being
questioned on the grounds that there was no evidence that it was
radiation (or E-M fields) that caused the miscarriages, as opposed to,
say, job stress. Unfortunately, I have not yet seen any kind of
overview of scientific opinions; I keep expecting something to appear
in _Science_. What I have seen and heard has been rather incomplete.
I don't know if George Byrne knows any more than the author of the
_New_Yorker_ articles chose to tell; if so, I wish he (or anyone else
who knows something) would tell us.

Alan McKenney E-mail: (INTERNET)
Courant Institute,NYU ...!cmcl2!robalo!mckenney (UUCP)


Date: Thu, 31 Aug 89 15:15:18 EMT
Subject: Positions at Bergen Scientific Centre


To emphasize its commitment to the environment IBM has appointed
its Bergen Scientific Centre -- as the centre for IBM's activities
within environmental modelling and as the focal point for information
on environmental activities throughout Europe.

In this regard the Bergen Scientific Centre has the following immediate job

Opportunity 1:

This position is to be filled by a PhD with a strong technical
background in Environmental Modelling, Remote Sensing, Databases,
Oceanographic Modelling, Mapping and Expert Systems. The successful
candidate should have research experience in at least two of these
areas and should have demonstrated research ability by publishing in
international refereed journals.

In addition the candidate must have good research experience in
parallel computing and must be able to show experience in leading
a scientific research team.

Opportunity 2:

This position is to be filled by a PhD with a strong research record
covering the area of Environmental Modelling including: pollution,
modelling of rivers, lakes and groundwater pollution, biological
modelling, deforestation and global warming.

In addition the candidate must possess a good publication record,
within international journals, and must have state-of-the-art
knowledge in parallel computing.

Bergen provides a pleasant environment in the heart of the Western
Fjords of Norway with excellent opportunities for outdoor pursuits
especially skiing, and water-sports. The centre is English speaking
and provides a friendly and flexible environment for scientists of
many different backgrounds.

Prospective candidates should send a detailed curriculum vitae to:

The Centre Manager, Bergen Scientific Centre, The High Technology Centre,
5008 Bergen, NORWAY.

Closing date for Applications: November 1, 1989.


From: Greg Ammar <>
Date: Thu, 31 Aug 89 16:30:09 EDT
Subject: Temporary Address Change for Ammar

Greg Ammar of the Department of Mathematical Sciences at Northern
Illinois University will be a Visiting Associate Professor in
the Department of Mathematics at the University of Kentucky
for the 1989-1990 academic year. His e-mail address until
June 1990 is "".


From: Pablo Jacovkis <>
Date: Fri, 1 Sep 89 17:29:54 ARG
Subject: Approximating Dirac Delta

I would like to have a good approximation of the constant c such that
the integral over Rn of the function

fi(x) = 0 if norm(x) >=1

fi(x) = c.exp(-1/(1-norm(x).norm(x))) if norm(x) <1

is 1.
c depends on the dimension n of the space Rn, and I am interested in the
cases n=1 and n=2, in order to approximate the Dirac delta with as smooth
a compact support mollifier as possible. It is easy to approximate the
delta with infinitely differentiable functions without compact support
or with functions with compact support but not infinitely differentiable;
I want both properties.

Pablo M. Jacovkis


From: Paul Calamai <>
Date: Sun, 3 Sep 89 16:24:44 EDT
Subject: Apollo Version of NAG Library

Perhaps someone out there in na land can help me with a
problem I'm having.

The University of Waterloo has obtained a site licence for
the NAG library. My research group has an Apollo DN3500
running OS SR10.1 with all three environments installed
(ie. bsd4.3, sys5, aegis) but with bsd4.3 as the default
SYSTYPE. The implementation of NAG we received contained a
compiled, tested, ready-to-use version of NAG Mark 13 Fortran
double-precision library for APOLLO DOMAIN hardware running
v9.7 AEGIS with fortran compiler v9.66 with the LBR utility.
We have v10.1 fortran. Because the v9.7 AEGIS implementation
is worthless under SR10.1 we have recompiled the sources twice
(once using options -compress -cpu 3000 and once using
options -dba -cpu 3000) and recreated the library
using ar. The problem is that revalidating the library
by running all the example problems identifies a number of
problems that fail and many (~20) that produce no output
beyond a header (identifying the routine) and the column
headings for a table that never gets printed.

I'm wondering if anyone else has gone through this exercise
with more success then we've had. Any recommendations
would be appreciated. If you need more details please
email me and I'll do my best to provide them.


Paul H. Calamai
Dept. of Systems Design Engineering
U of Waterloo
(519)885-1211 x3182


End of NA Digest