NA Digest Sunday, May 14, 1989 Volume 89 : Issue 19

Today's Editor: Cleve Moler

Today's Topics:


From: Ronald England <>
Date: 9-MAY-1989 14:41:47 GMT
Subject: London Conference on Ordinary Differential Equations

Following various comments which have been circulating in connection with
the IMA Conference on Computational Ordinary Differential Equations,
I wish to circulate, as widely as possible, the following statement
on behalf of the organising committee, together with an updated
version of the conference application form.

We were well aware, from an early stage, that London would be an
expensive location in which to organise a conference. Imperial College
insisted upon charging the IMA at commercial rates, and we were
unsuccessful in our requests for external funding. We have also made
arrangements for what we think is an impressive list of invited and
highlighted speakers, all of whom require some financial support.

Nevertheless, when the IMA officers calculated and announced the level
of fees for the conference, we were greatly disturbed to discover that
these were far in excess of that which we had been expecting.

You will probably have read the justification of those fees which has
been circulated by Catherine Richards, Secretary and Registrar of the IMA,
and which was accompanied by an announcement of a new, reduced level of fees.

While these fees are still higher than we would have hoped for,
and while we feel that they should not be taken as a precedent by
organisers of future conferences when calculating their costs and fees,
we must nevertheless express our gratitude to the IMA officers and
others who have made an effort to bring down the costs of the meeting.
In spite of some of the unfortunate comments which have been made, we
hope that the conference will now be able to go ahead, and that, with
the support of all the interested community, it will be as successful
an event as any of the previous conferences on a similar subject area.

May I also take this opportunity of reminding you that this meeting
will take place in the week following the regular Numerical Analysis
Conference at the University of Dundee. Those of you who may
be coming from outside Great Britain will therefore be able to
share you travelling expenses between the two conferences.

We look forward to seeing as many of you as possible in London.

Roland England


Conference on

Imperial College, London 3rd-7th July, 1989

The IMA is organising a conference on Computational Ordinary Differential
Equations, to be held at Imperial College, London, in July, 1989.

The aim of this meeting is to bring together research workers and
practitioners in theory and algorithms for the numerical solution
of ordinary differential equations . All relevant topics are to be
covered and contributions in application areas will be encouraged .

The invited speakers are:

# J.C. Butcher (Auckland)
# H.B. Keller (Caltech)
# R.E. O'Malley Jr. (Rensselaer Polytech. Inst.)
# L.R. Petzold (Lawrence Livermore)
# R.D. Russell (Simon Fraser, Vancouver)
# J.M. Sanz-Serna (Valladolid)
# L.F. Shampine (Southern Methodist, Dallas)
# M.N. Spijker (Leiden)

Talks by the following speakers will also be highlighted:

# U. Ascher (British Columbia)
# G. Bader (Heidelberg)
# G.D. Byrne (Exxon Research)
# G.F. Corliss (Milwaukee)
# U. Kirchgraber (ETH, Zurich)
# J.F.B.M. Kraaijevanger (Leiden)
# J. Lawson and M. Berzins (Leeds)
# B.J. Leimkuhler and O. Nevanlinna (Helsinki)
# H.W. Tam (Illinois)

The programme will also include quite a number of
shorter 20 minute contributed papers , and posters .

An optional visit to Hampton Court Palace will be organised on the
afternoon of Wednesday, 5th July . This afternoon is free from lectures .

Members of the Organising Committee for the conference are:
Dr. J.R. Cash (Imperial College of Science and Technology, London),
Dr. R. England, AFIMA (The Open University, Milton Keynes),
Prof. I. Gladwell, FIMA (Southern Methodist University, Dallas) and
Dr. A. Iserles, AFIMA (University of Cambridge).

Authors who have been notified that their papers have been accepted for
presentation are now invited to submit them for publication in the
proceedings , which are to be published by Oxford University Press in the
IMA Conference Proceedings Series . In order to expedite the refereeing
procedure , full papers should be submitted before the end of the meeting .

FEES in Pounds Sterling (including VAT at 15%) VAT Reg. No. 250 4466 77
Members Non-Members Bona-fide Bona-fide
Postgraduate Postgraduate
Student Student
Conference fee Members Non-Members
(including coffee, lunch and
tea on 3rd-7th July, inclusive) 115.00 145.00 85.00 100.00

Residential fee
(including dinner, bed and
breakfast at Imperial College
in single study bedrooms for
the nights of 2nd-6th July 120.00 120.00
inclusive, and conference
dinner on 6th July
-Alternatively please write
for a list of local hotels

Non-residents only
Conference dinner on 6th July 30.00 30.00

Visit to Hampton Court Palace
(including entrance to Palace
and Maze and coach) 8.00 8.00

Late registration fee
Applications received
later than 2nd June 10.00 10.00
Applications received
later than 23rd June 15.00 15.00

Invoice fee 10.00 10.00


To: Miss Yvonne May, Conference Officer
The Institute of Mathematics and its Applications
Maitland House, Warrior Square
Southend-on-Sea, Essex SS1 2JY, England


Imperial College, London 3rd-7th July, 1989









Please charge my Access Account Number
If your Access card account address is different from the above address ,
please enter here:

All cheques should be made payable to
The Institute of Mathematics and its Applications.


From: Hans Stetter <E115N06%AWITUW01.BITNET@Forsythe.Stanford.EDU>
Date: 10 MAY 89 15:53:37
Subject: London Conference on Computational ODE

From Ronald England`s announcement in this na-digest you will have seen
that the Organizing Committee of the Computational ODE Conference has finally
decided to hold the conference as planned. Obviously, their decision was moti-
vated not so much by the reduction of the conference fee which has been ach-
ieved but by the consideration of the huge effort which has been invested in
this conference by so many people.

Since I have served as the catalyst who brought to light the dissatis-
faction of a great many of you, I feel that I should also speak up at this
point. Naturally, I cannot claim that the complaints which I had voiced have
become immaterial now. However - putting myself in the position of a member
of the Organising Committee in this moment - I must admit that I would have
reached the same conclusion: It would simply have been grossly unfair to all
those who have spent long hours in preparing their talks for us to cancel this
meeting two months before its scheduled time. Also, a good number of you will
have committed themselves by the purchase of non-refundable reduced airline
tickets and the like.

Besides, I am certain that this will be an outstanding meeting from the
scientific point of view. The selection of the invited and highlighted speak-
ers is impressive, and there will be many more highly stimulating contributed
papers - judging from the few that have come to my knowledge. All that remains
in order to make this conference a full scientific success is our own parti-

Therefore, I have decided to go to London for the conference - and I would
like to urge all of you who work in computational ODEs to try hard to be there.
Scientific meetings consist not just of the papers which are delivered (and
which could also be read at home) but of the colleagues and friends you meet
and of the information you get in conversation. And this cannot be had by
staying away!

Finally, a word of thanks to all who have responded to my previous ap-
peals in one way or other; excuse me for not being able to answer each of you
personally. I am quite sure that this example will not go unnoticed, and you
have contributed to its (partial) success.

See you in London! So long, Hans J. Stetter.


From: Chris Henrich <>
Date: 5 May 89 19:00:17 GMT
Subject: Test Data for Mathematical Functions

I would like to get my hands on data to help me test the accuracy
of mathematical functions. The functions in question are
"elementary" - square root, exponential, logarithm, direct & inverse
trigonometric, etc. The data that I want are pairs (x, f(x)) in
IEEE floating-point formats, both single and double precision.

I have already looked at one large catalogue of microcomputer software,
and drawn a blank. Commercially marketed stuff seems to be programs,
which actually *do* interesting stuff. What I am after is a bunch of
numbers that just sit there.

Any pointers toward such data, or toward catalogues of resources like
that, will be gratefully received.

Perkin-Elmer DSG, Tinton Falls, N.J.

UUCP: ...!rutgers!petsd!cjh
(201)758-7288 106 Apple Street, Tinton Falls,N.J. 07724


From: Jack Dongarra <>
Date: Thu, 11 May 89 17:19:44 CDT
Subject: Fourth SIAM Conference on Parallel Processing for Scientific Computing

Title: Fourth SIAM Conference on Parallel Processing
for Scientific Computing
Place: Chicago, Illinois

Conference Date: December 11-13, 1989
Abstract Deadline: June 1, 1989

Contact: SIAM Conference Coordinator
117 S. 17th Street, 14th Floor
Philadelphia, PA 19103-5052 USA
FAX: 215-564-4174

Purpose: The conference will focus on the development of analytic methods,
efficient algorithms, and applications software in context with
advances in computer architecture as applied to high-performance
systems. SIAM has selected these areas for a conference because
of the growing recognition of the importance of exploiting massive
parallelism to attack complex problems. In addition to new
computational algorithms there is a tremendous need for tools to
aid the development of programs for these new high performance
computers and also the need to develop new tools to aid scientists
in the visualization and analysis of their data. SIAM believes that
the development of effective large-scale high-performance scientific
computing systems depends, at least in part, on the continued
interaction of algorithm developers, numerical analysts, computer
architects, and scientists concerned with scientific computing.

Topics: Massively parallel computing, visualization of scientific
computation, tools for parallel algorithm development, and
many other related subjects.

Speakers: Alan Egolf, United Technologies Research Center
Dennis B. Gannon, Indiana University, Bloomington
David Gelernter, Yale University.
William Goddard, California Institute of Technology
W. Danny Hillis, The Thinking Machines Corporation
Ken Kennedy, Rice University
James Sethian, University of California, Berkeley
Andries van Dam, Brown University
Karl-Heinz A. Winkler, Los Alamos National Laboratory

How to Register: Advance registration material will be available in
late September 1989. To ensure receiving your registration
material and preliminary program, please send a request to or call: (215) 564-2929.

Contributed and poster presentations are invited in all areas consistent
with the conference themes. A description of your presentation must be
submitted on a SIAM abstract form, which can be obtained by contacting SIAM.

A contributed presentation will be considered for one of the following
categories, a 30-minute talk, a 20-minute talk or a poster presentation.
When submitting your abstract, please indicate the type of presentation
for which you would like to be considered. A limited number of papers
will be selected for 30-minute talks, but all others will be considered
for 20-minute talks or poster presentations. In addition, we also ask
that you submit an extended abstract, typed on plain white bond, of 750
words and also an abstract of 100 words submitted on a SIAM Abstract
Form. The reason for this is that the 100 word abstract will be used in
the description of your talk in the program and the 750 word abstract
will be used in selecting papers to be presented at the conference. The
selection of papers will be made in the first week of July 1989.

The conference organizers have set aside exclusive times and meeting
rooms for vendors to give brief presentations about their products and
services. A brief description of your talk, not exceeding 100 words,
must be submitted on a SIAM abstract form, which can be obtained by
completing the attached postage-paid reply card and returning it to
SIAM. [Ed. Note-- Attached to what, Jack?] A presentation consists
of a seventeen-minute talk, followed by three minutes for questions.


From: Steve Snelgrove <...!wicat!steves>
Date: 11 May 89 21:38:30 GMT
Subject: Validation of Floating Point Operations in C

I am trying to find a C compiler for the IBM PC which
supports IEEE floating point. Surprisingly (at least to me),
this has turned out to be a very unsuccessful effort. While many
compilers allow access to a coprocessor via the hardware floating
point instructions, the corresponding software routines do not
produce the same results.

The tool I am using in my investigations is a C program
which performs arithmetic according to the standard. I have
implemented four operations on the type 'double': add, subtract,
multiply, and divide. The test driver portion of the program
uses an array containing about 100 entries of both positive and
negative values representing zeros, subnormals, normals, infinities,
and NaN's. Two nested 'for' loops are used to generate the power set
of these elements. For each pair of values, an operation is performed
twice, once using the simulation and then again using the intrinsic
operation. The results are then compared for equality. Command line
options control the selection of tests and the printing of results.
For each operation, a summary report is displayed which should be of
the form:

ZERO | 100% | 100% | 100% | 100% | 100% |
SUBNORMAL | 100% | 100% | 100% | 100% | 100% |
NORMAL | 100% | 100% | 100% | 100% | 100% |
INFINITY | 100% | 100% | 100% | 100% | 100% |
NAN | 100% | 100% | 100% | 100% | 100% |

The test shown above was generated by a 68881 coprocessor. I
have at this location a 68000/68881 machine and several IBM machines,
one of which has a 80287. I have tried an early version of the simulation
on an 80387. All of the coprocessor results seem to agree at least with
regards to my test values. Two exceptions are experienced as a result
of incorrect initialization and interrupt support when using either an
8087 or an 80287.

It seems to me that the C startup code should set the infinity
mode on 8087's and 80287's to affine mode rather than the default which is
projective. This will cause these chips to perform according to the
standard and in the same manner as an 80387 which only supports affine mode.

I also think that the C startup code should supply an interrupt
handler which will normalize subnormal (denormal) operands. I think that
this will prevent the generation unnormalized results in multiply and
divide. The storing of unnormalized values causes an invalid operation
exception. The IEEE standard requires normalizing mode.

I believe that these two steps are required in the startup
code. I have worked with both the Lattice and the Microsoft C compilers
and have not found these steps to be present.

This does not surprise me too much in the case of Lattice since
their only support of the standard is in the use of the storage formats.
All underflows are forced to zero, infinities and NaN's are not recognized,
and subnormals cause 'printf' to go into an infinite loop.

Microsoft does the best job I have seen for the IBM PC. Still
they do some funny things like force the sign of zero to be positive
(but only after returning from a subroutine?). The standard indicates
that the default mode of operation should be with exceptions masked.
Microsoft requires a call to a 'control87' function to mask exceptions.
I also experienced a problem when multiplying a subnormal and a normal
resulting in invalid as described above. The software emulation correctly
produced a result. However, when run on the machine with the 80287 the
exception was not masked and caused execution to terminate with an error
message. Another problem was observed in operations involving two NaN's.
Apparently an attempt was made to emulate the operation of the coprocessor
which returns the NaN with the larger magnitude. However, the comparison
was made from LSW to MSW instead of from MSW to LSW.

As a side note on operations with NaN's, the 80387 manual seems
to indicate that Signaling NaN's will be converted to Quiet NaN's by
setting the high order bit in the fraction. I did not see this on the
one 80387 I tried. Are there differing versions of the 80387 or maybe
I goofed by not having enough NaN cases when I did the test?

Another C related issue which caused me considerable confusion
and delay is that of the double rounding which can occur when results
are first rounded to the size of the internal registers (extended) and
then rounded again when storing to a double. The only mention of this
that I can recall having seen is in the Apple Numerics Manual. On page
38 of the second edition, they state:

| The IEEE Standard encourages extended-precision expression evaluation.
| on rare occasions, extended evaluation produces results slightly
| different from those produced by other IEEE implementations that lack
| extended evaluation. Thus, in a single-only IEEE implementation,
| z := x + y;
| with x, y, and z all single, is evaluated in one single precision
| operation, with at most one rounding error. Under extended evaluation,
| however, the addition of x + y is performed in extended; then the result
| is coerced to the single precision of z, with at most two rounding errors.
| Both implementations conform to the standard.

This seems to directly contradict the wording of the standard which says
in section 4.3 that:

| ..., a system that delivers result only to double extended destinations
| shall permit the user to specify rounding to single or double precision.
| Note that to meet the specifications in 4.1, the result cannot suffer
| more than one rounding error.

How is Apple justified in their statement made above?

Apple seems to be the only company which has really tried to
provide full IEEE floating point support to the users of high level
languages. What kind of feelings do people hold about the success
of their efforts? Have they adequately covered all of the relevant
issues with regards to users requirements for control and usage of
the floating point environment? Is there anything as well defined
available for other machines?

My test program has been posted to the newsgroup
sci.math.num-analysis. If you don't have access to that network,
and would like a copy, please contact me directly.

Steve L. Snelgrove
WICAT Systems
Research and Development
748 N. 1340 W.
Orem, UT 84057


From: Rex Jaeschke <aussie!>
Date: 13 May 89 23:49:10 GMT
Subject: Numerical C Extensions Group (NCEG) Meeting Report

The following report is reprinted from Volume 1, Number 1 (June 1989)
of The Journal of C Language Translation, Copyright 1989 Rex
Jaeschke. Permission is granted for duplication and distribution of
this report for the purposes of furthering the work of NCEG.

Numerical C Extensions Group Status

Rex Jaeschke
NCEG Convener


When I conjured up the idea for an ad hoc group to define numerical
extensions to C earlier this year, I had no idea as to what the
reaction would be. The evidence is now clear that this endeavor is
seen as being very worthwhile. Not only have more than 90 people
asked to be added to the contact database, but 30 of them attended
the one-and-a-half day meeting at Cray Research on May 10--11.

The backgrounds of the attendees was diverse. The supercomputing
industry was represented via Cray, Convex, Supercomputer Systems, and
Thinking Machines. The IEEE community was well represented by Hough
(from Sun), Cody (from Argonne Labs), and Thomas (from Apple.) Other
organizations represented included Unisys, Microsoft, Digital
Equipment Corporation, H-P, CDC, IBM, Solborne, Farance, Inc.,
University of Minnesota, Intermetrics, and Information and Graphics
Systems. The digital signal processing industry was represented by
Analog Devices, and LLNL, Army BRL, and Polaroid Corporation
represented the user community. Dennis Ritchie from AT&T also

There was no real sentiment that we deliberately go against the
direction established by ANSI C. In fact, quite the contrary.
However, it was recognized that some of ANSI C's constraints may
impede our activities resulting in possible conflicts. The whole
issue of errno and formatted I/O of NaNs and infinity are examples.

The Issues

The main purpose of the meeting was to identify and prioritize the
principal technical issues. The group then voted on each topic
indicating high or medium (or no) priority. The high priority votes
were weighted twice as much as the medium, and the following list of
priorities resulted.

Main Numerical Issues
Topic Priority
aliasing 29
vectorization 27
complex 27
variably dim arrays 25
IEEE issues 24
exceptions/errno 24
float/long double library 23
parallelization 22
ANSI <math.h> 21
array syntax 19
extra math functions 17
aggregate initializers 15
inter-language issues 15
wide accumulators 10
math function precision 9
non-zero-based arrays 8
numerical representation 6
new data types 4
new operators 4
function overloading 4

Another topic, ``Arrays as first class objects'' had a high priority
(21) but after considerable debate was dropped from the list since it
was agreed its addition would likely cause great confusion to existing
C programmers.

Formation of Subgroups

The bulk of the agenda time was then given to the top ten topics, each
getting 20--30 minutes. For each of these topics, attendees
volunteered to be the primary and alternate coordinator. (The minutes
of the first meeting identify these people. In the interim, contact me
for details.)

The intent is that the real technical work will go on between meetings
and be coordinated by the leaders of each subgroup. Then, at the
following meeting, each subgroup will present the results of its work
and make formal proposals as appropriate. This way, the committee can
focus on the final, distilled issues rather than everyone getting
involved at all levels. It will also significantly reduce the amount
of paper in the mailings.

If you wish to participate in any of these subgroups it is your
responsibility to contact the leaders and identify yourself, your
concerns and how you can help. If your area of interest is not listed
here, start your own subgroup and let me know.

Mailings and Submissions

Most of people interested in NCEG appear to have an e-mail address so
that should make the subgroups' job much easier in coordinating
various viewpoints and proposals. However, all formal distributions
will be by paper mail. Since meetings are to be once every six months
there will be two mailings between meetings. The first will occur
within 4--6 weeks after a meeting and will contain minutes, new papers
and other appropriate correspondence. The second will occur about 4--6
weeks prior to the following meeting. The cut-off date for formal
submissions for the September meeting is August 11.

Forward all correspondence to me (either by mail or via
uunet!aussie!rex) and I will assign it a document number.
(Note that I do not have a troff formatter.) However, do that
only if your paper is concerned with issues other than those being
handled by the subgroups. For subgroup issues, forward papers to the
subgroup coordinators so they can include it in their submissions to
me. The intent is to avoid excessive duplication of points and to
allow the short meeting time to be used more effectively. The more
formal documents we have the slower it will go.

Tom MacDonald at Cray Research has agreed to do the mailings, at
least for the interim. Frank Farance of Farance, Inc., has
volunteered to be the redactor of the group's working document.
Thanks to Tom and Frank. (Thanks also to Randy Meyers from DEC, who
acted as meeting secretary and to Cray for being meeting host.)

Formal Affiliation

There was general consensus that we become affiliated with a
recognized standards organization. The final proposal was that we
become a working group within X3J11. If we follow that route, it will
result in our publishing a Technical Report, a non-binding report on
our findings and recommendations. With suitable planning, we might be
able to have that elevated to a Technical Bulletin and get it
distributed with the ANSI Standard. Getting our extensions adopted as
a standard is also possible, in the long term. At this stage, I plan
to ask for agenda time at the next X3J11 meeting to discuss admitting
us as a work group.

In the interest of economy, the next two meetings are scheduled in the
same location and week as those of ANSI C's X3J11. These NCEG meeting
dates are September 19--20 (Salt Lake City, Utah), and March 7--8,
1990 (New York City.)

Rex Jaeschke | C Users Journal | Journal of C Language Translation
(703) 860-0091 | DEC PROFESSIONAL |1810 Michael Faraday Drive, Suite 101
uunet!aussie!rex | Programmers Journal | Reston, Virginia 22090, USA


From: Kathleen LeBlanc <>
Date: Mon, 8 May 89 11:31 EDT
Subject: New Books From SIAM

TITLE: Mathematical Aspects of Vortex Dynamics
SERIES: Proceedings in Applied Mathematics
EDITOR: Russel E. Caflisch
Courant Institute of Mathematical Sciences, New York University

224 pages, softcover, ISBN 0-89871-235-1, April 1989
LIST PRICE: $29.50

This volume presents 19 papers representing the invited and contributed talks
at the workshop on Mathematical Aspects of Vortex Dynamics, held at the Xerox
Training Center in Leesburg, Virginia, on April 25-27, 1988.

The book discusses a variety of fluid problems, including vortex sheets and
shear layers, the vortex reconnection problem, vortex methods, coherent vortex
structures, fluid dynamic stability, systems of point vortices, turbulence
modeling and other applications. It is written for students and researchers
working in the area of fluid dynamics, including mathematicians, engineers and
physicists. Anyone wishing to survey the current state of the subject of
vortex dynamics is especially encouraged to order this book. Knowledge of
fluid dynamics is a recommended prerequisite.

CONTENTS: Mathematical Analysis of Vortex Dynamics, Russel E. Caflisch;
Improved Vortex Methods for Three-Dimensional Flows, G. Winckelmans and A.
Leonard; The Relation Between Thin Vortex Layers and Vortex Sheets, M.J.
Shelley and G.R. Baker; Computations of Broad Band Instabilities in a Class of
Closed-Streamline Flows, B.J. Bayly; Vortex-Sheet Dynamics and Some References
to Hyperfunction Theory, T. Kambe; A Free Surface Vortex Method with Weak
Viscous Effects, T.S. Lundgren; An Iterative Method for Computing Steady
Vortex Flow Systems, Bruce Turkington and Alexander Eydeland; On Invariant
Measures for the 2-D Euler Flow, Mario Pulvirenti; On Similarity Flows
Containing Two-Branched Vortex Sheets, D.I. Pullin; Strain-Induced Vortex
Stripping, David G. Dritschel; Convergence of the Vortex Method for Vortex
Sheets, John S. Lowengrub; Boundary Conditions and Deterministic Vortex
Methods for the Navier-Stokes Equations, Georges-Henri Cottet; Observations on
Vorticity Creation Boundary Conditions, Christopher R. Anderson; Vortex
Dynamics of Stratified Flows, Gretar Tryggvason; Remarks on Vortex Breakdown,
Mel S. Berger; Numerical Studies of Vortex Reconnection, D.I. Meiron, M.J.
Shelley, W.T. Ashurst, and S.A. Orszag; Vortex Lattices in Theory and
Practice, Laurence J. Campbell; The dynamics of Vortex Structures in the Wall
Region of a Turbulent Boundary Layer, John L. Lumley; The Energy of a Vortex
Lattice Configuration, Kevin A. O'Neill


TITLE: Combinatorial Algorithms: An Update
SERIES: CBMS-NSF Regional Conference Series in Applied Mathematics
EDITOR: Herbert S. Wilf, Professor of Mathematics, University of Pennsylvania

56 pages, softcover, ISBN 0-89871-231-9, April 1989
LIST PRICE: $12.00

Combinatorial Algorithms: An Update focuses on some of the exciting, vigorous
new work being done in this field, including recent developments in algorithms
for generating combinatorial objects such as partitions, Gray codes, and

trees. This monograph is a comprehensive update of Combinatorial Algorithms
by Wilf and Albert Nijenhuis, which was first published in 1975 and revised in
1978. It is based on a series of ten lectures given by Wilf at the CBMS-NSF
Conference on Selection Algorithms for Combinatorial Objects held in 1987 at
Colorado College.

This volume presents a survey of some of the new work that has been done since
the appearance of the second edition of Combinatorial Algorithms. Topics that
are discussed include progress in the following areas: Gray Codes, listing of
subsets of given size of a given universe, listing rooted and free trees,
selecting free trees and unlabeled graphs uniformly at random, and ranking and
unranking problems on unlabeled trees.

This book is excellent for graduate students, researchers, and computer
science professionals in discrete mathematics, with the only prerequisite
being the successful completion of a high-level undergraduate course in

CONTENTS: The Original Gray Code; Other Gray Codes; Variations on the Theme;
Choosing 2-Samples; Listing Rooted Trees; Random Selection of Free Trees;
Listing Free Trees; Generating Random Graphs; Bibliography


TITLE: Stochastic Processes in the Neurosciences
EDITOR: Henry C. Tuckwell
SERIES: CBMS-NSF Regional Conference Series in Applied Mathematics No. 56

136 pages, softcover, ISBN 0-89871-232-7, May 1989
LIST PRICE: $24.50

This monograph is centered on the quantitative analysis of nerve-cell
behavior. The work is foundational, with many higher order problems still
remaining, especially in connection with neural networks. Thoroughly
addressed topics include stochastic problems in neurobiology, and the
treatment of the theory of related Markov processes.

Special features include systems of nonlinear stochastic partial differential
equations and the pertubative approach to their solutions; comprehensive
treatment of channel noise; and statistical analysis of point processes;
Wiener kernel explanations.

Mathematicians, especially probabilists and statisticians interested in new
applications, will appreciate Tuckwell's work, as will neurobiologists.
Background requirements include intermediate probability and related
mathematics, as well as a rudimentary knowledge of neurophysiology.

CONTENTS: Deterministic Theories and Stochastic Phenomena in Neurobiology;
Synaptic Transmission; Early Stochastic Models for Neuronal Activity Including
Poisson Processes and Random Walks; Discontinuous Markov Processes with
Exponential Decay; One-dimensional Diffusion Processes; Stochastic Partial
Differential Equations; The Statistical Analysis of Stochastic Neural
Activity; Channel Noise; Wiener Kernel Expansions; The Stochastic Activity of
Neuronal Populations


TITLE: Parallel Processing and Medium Scale Multiprocessors
SERIES: Proceedings in Applied Mathematics
EDITOR: Arthur Wouk, Mathematical Sciences Division, Army Research Office

216 pages, softcover, ISBN 0-89871-238-6, June 1989
LIST PRICE: $31.50

The papers in this book represent those presented at the ARO research workshop
on Parallel Processing and Medium Scale Multiprocessors held at Stanford
University in January 1986. The workshop brought together research workers
with strong connections to scientific computation as well as an interest in
the systems programming problems that are restricting the application of the
new hardware being developed.

This volume contains papers covering several areas, including systems
programming, parallel language/programming systems, and applications
programming. The work reported includes investigations into debugging of
operating systems, portability of applications programs and parallel operating
systems, efficient resource allocation in multiprocessors, parallel constructs
for applications programming languages, and applications programming efforts
on diverse architectures. These papers can provide helpful guidance for those
wishing to explore the new territory of automatic optimal parallelization of
sequentially conceived programs.

This book will be a useful guide for all researchers having an interest in
scientific computation, especially those wishing to explore more thoroughly
the problems involved in automatic optimization in the compilation of parallel

CONTENTS: Debugging Multi-Task Programs, Morven Gentleman and Darlene
Stewart; Matrix Computations and Game Playing on the iPSC, Cleve Moler and
David Scott; A System for Parallel Processing, P. O'Leary, G.W. Stewart, and
Robert van de Geijn; Statistical Methodologies for the Control of Dynamic
Remapping, Joel Saltz and David Nicol; Sparse Cholesky Factorization on a
Local-Memory Multiprocessor, Alan George, Michael Heath, Joseph Liu and Esmond
Ng; Concurrent Global Optimization on a Network of Computers, Richard Byrd,
Cornelius Dert, Alexander Rinooy Kan, and Robert Schnabel; Heterogeneous
Processes on Homogeneous Procesors, George Cybenko, David Krumme, K.N.
Venkataraman and A. Couch; Matrix Multiplication on Boolean Cubes using
Generic Communication Primitives, Lennart Johnsson and Ching-Tien Ho; The
Force on the Flex: Global Parallelism and Portability, Harry Jordan; SCHEDULE:
An Aid to Programming Explicitly Parallel Algorithms in Fortran, J. Dongarra
and D.C. Sorenson; Dynamic Grid Manipulation for PDEs on Hypercube Parallel
Processors, William D. Gropp; Solving Compressible Euler Equations on a
Hypercube Simulator, Jung Pyo Hong, Bob Tomlinson, and Nisheeth Patel


LE BLANC AT SIAM. Phone: (215)564-2929; E-mail:


End of NA Digest