NA Digest Sunday, March 12, 1989 Volume 89 : Issue 10

Today's Editor: Cleve Moler

Today's Topics:


From: John Larson <>
Date: Mon, 6 Mar 89 11:22:48 CST
Subject: NCSA (Univ. Illinois) Conference on Parallel and Vector Processing

NCSA Second Conference on Parallel
and Vector Processing May 8-10, 1989

The goal of the NCSA Second Conference
on Parallel and Vector Processing is to
provide information to the participants on
the latest developments in parallel and
vector architecture, applications, algorithms,
performance, and programming environments.

Monday, May 8, 1989

Keynote Address David Kuck, CSRD

CRAY-2 Robert Numrich, CRI
CRAY Y-MP Ram Gupta, CRI
ETA10 Cliff Arnold, ETA
Myrias Martin Walker, Myrias
CEDAR Kyle Gallivan, CSRD
NCUBE Doup Harless, NCUBE
CM-2 Jill Mesirov, Thinking Machines

Visualization Theatre Maxine Brown, UIC

Tuesday, May 9, 1989

Performance Evaluation
Perfect Club Michael Berry, CSRD

QCD Dennis Duke, FSU
Device simulation Karl Hess, CSL-UI
CFD Karl-Heinz Winkler, LANL
Weather modelling Robert Wilhelmson, NCSA
Biology Michael Ess, Intel
Chemistry Jan Andzelm, CRI

Visualization Theatre Donna Cox, NCSA

Wednesday, May 10, 1989

Programmer's Environment
Parallel Computing Forum Bruce Leasure, KAI
Autotasking Mark Furtney, CRI
Development environment Daniel Reed, DCL-UI

Numerical Algorithms
LAPACK Jack Dongarra, Argonne
Multitasked libraries Qasim Sheikh, CRI
Algorithm development Ahmed Sameh, CSRD
Matrix solvers on CM-2 Creon Levit, NASA
Algorithms for Transputers Ron Cok, Kodak

For additional information
Call Michael Welge, Manager of the parallel
processing program at NCSA (217) 244-1999
or email (Internet)
or 13016@ncsavmsa (Bitnet)


From: Jack Dongarra <>
Date: Thu, 9 Mar 89 16:18:06 CST
Subject: PDE position at Argonne

The Mathematics and Computer Science (MCS) Division of Argonne
National Laboratory invites applications for a regular staff
position in the area of advanced scientific computing, with emphasis
on the numerical solution of partial differential equations.

Applicants with a Ph.D. in (applied) mathematics or computer
science will be given preference; however outstanding candidates
with degrees from other disciplines will be considered. The
position requires extensive knowledge of numerical methods for
partial differential equations, research experience in at least
one application area, and a strong interest in advanced (parallel)
architectures and state-of-the-art visualization techniques. Several
years of research experience beyond the doctorate are desirable, as is
familiarity with advanced architectures and visualization techniques.
Applicants must have an established record of research accomplishments,
as evidenced by publications in refereed journals and conference

The MCS Division offers a stimulating environment for basic
research. Current research programs cover areas of applied
analysis, computational mathematics, and software engineering, with
emphasis on advanced scientific computing. The division operates
the Advanced Computing Research Facility (ACRF), which comprises a
network of advanced-architecture computers, ranging from an
8-processor Alliant FX/8 to a 16,384-processor Connection Machine
CM-2, and a graphics laboratory. A network of Sun workstations
supports the general computing needs of the division.
Argonne's central computing facilities include a CRAY X/MP-14;
additional access to supercomputers is provided through the major

Argonne is a multipurpose national laboratory operated by the
University of Chicago for the U.S. Department of Energy. It is
located about 25 miles southwest of Chicago.

For consideration, send detailed resume to Rosalie L. Bottino,
Employment and Placement, Box J-MCS-37017-83, Argonne National
Laboratory, 9700 S. Cass Avenue, Argonne, IL 60439. For more
technical information, contact Dr. Hans G. Kaper, Director, MCS
Division at 312-972-7162 ( Argonne is an equal
opportunity/affirmative action employer. Women and minorities are
especially encouraged to apply.

Applications will be considered until the position is filled.


From: David Bailey <>
Date: Mon, 13 Feb 89 08:09:58 PST
Subject: Euler's constant

I have computed Euler's constant to high precision in conjunction with
some studies of possible interrelationships between fundamental
constants of mathematics, using Ferguson's algorithm. The method I
used was described in my paper "Numerical Results on the Transcendence
of Constants Involving Pi, E, and Gamma", Mathematics of Computation,
Vol. 20, No. 181 (January 1988), p. 275-281. I also have a more
recent paper on the subject that is due to appear in Mathematics of
Computation later this year. If you do not have access to MOC, let me
know and I will send you copies.

The scheme is basically the formulas

inf m
2^n --- 2^{mn} --- 1
gamma = -------- \ -------- \ ----- - n log 2 + O(2^{-n} e^{-2^n})
e^{2^n} / (m+1)! / t+1
--- ---
m=0 t=0

--- 1
log 2 = \ ---------------
/ (2k-1) 3^{2k-1}

Using these formulas, the value of gamma to 180 decimal places is

10 ^ -1 x 5.77215664901532860606512090082402431042159335939923598805

David H. Bailey
Mail Stop 258-5
NASA Ames Research Center
Moffett Field, CA 94035
Telephone: 415-694-4410


From: Murli Gupta <MMG%GWUVM.BITNET@Forsythe.Stanford.EDU>
Date: Wed, 8 Mar 1989 13:51 EST
Subject: More on Kerner's Method for Polynomial Root Finding

In NA Digest <Jan 29, 1989 Vol 89:No.4>, Lee Dickey asked about Kerner's
method. A new book just landed on my desk that contains a reference to this
method. The book is: Precise Numerical Analysis by Oliver Aberth,
W.C. Brown Publ., 1988.
This is the first book I have found to contain a reference to Kerner.
I quote from page 91:
The method of refining zero approximations by formula (6.30) was
discovered independently by Durand,E. [Solutions Numeriques des
Equations Algebriques, Tome 1, Equations du type F(x)=0. Racines
d'un Polynome, Masson, Paris, 1960, 277-280] and Kerner, I.O.
[Numer. Math. 18(1966), 290-294]. The formula (6.30) can also be
used to obtain zero approximations [Aberth, O., Math. Comp. 27(1963)
339-344], but this is not as efficient as the other methods given
in this chapter.

Kerner's paper was reviewed by J.F. Traub in Math Rev.: MR34 #3778.
Another of her paper appeared in Z.A.M.M. Vol 47 (1967), pp 549-550
and was reviewed by H.E. Fettis in MR 39 #3696.
Her Ph.D. thesis(1961) was reviewed by G.Meinardus in MR 32 #2801.

Murli Gupta 202/994-4857
Department of Mathematics mmg@gwuvm.bitnet
George Washington University, Washington, D.C. 20052


From: Kaj Madsen <>
Date: Mon, 13 Mar 89 10:32:41 +0100
Subject: Durand-Kerner's Method.

January 27 L.J.Dickey requested information on 'Kerner's Method'. Since
the method may be of general interest I send this message to the net.
First of all, Durand actually introduced 'Kerner's Method' six years
before the paper by Kerner appeared and more information can be found
in the paper by G.Kjellberg (BIT 24:4 1984) and the one by H.Guggenheimer
(BIT 26:4 1986).

The resemblance with Newton's Method is easily explained: It IS Newton's
Method applied to the non-linear system which describes the roots in terms
of the coeffients of the polynomial.

Joergen Sand,DIKU,Copenhagen (using the adress of na.madsen).


From: Samir Chettri <>
Date: 8 Mar 89 19:18:06 GMT
Subject: Kalman Filtering and Quality Control

I have been trying to find out if any work has been done in applying
the Kalman Filter to the Statistical Quality Control Problem at all.
If so, are there any references, books etc. that are available ???

Also on a related note, what is the text/paper that gives a good
and clear exposition on the Kalman Filter especially from the
Multivariate Statistical/Least Squares view point ??


Samir Chettri (


From: Luciano Molinari <>
Date: 10 Mar 89 11:24 +0100
Subject: Chaos on PC's

Does anybody know anything about Chaos theory demonstration programs
for MS-DOS PC's?
Thanks for helping,
Luciano Molinari.


End of NA Digest