NA Digest Sunday, February 12, 1989 Volume 89 : Issue 6

Today's Editor: Cleve Moler

Today's Topics:


From: Steve McCormick <smccormi%copper@boulder.Colorado.EDU>
Date: Mon, 6 Feb 89 14:54:47 mst
Subject: Copper Mountain Multigrid Conference

Fourth Copper Mountain Conference
April 9-14, 1989

Organized by: The University of Colorado at Denver
The Center for Applied Parallel Processing
The Society for Industrial and Applied Mathematics

Sponsored by: The United States Air Force Office of Scientific Research
The Gesellschaft fuer Mathematik und Datenverarbeitung


Multigrid Tutorial
Advanced Workshop
Parallel Multigrid Panel

Novel Parallel Methods

Chairmen: Jan Mandel and Steve McCormick, CU-Denver

Chairmen: Joel Dendy, Los Alamos National Lab.
Seymour Parter, Univ. Wisconsin
John Ruge, CU-Denver
Klaus Stueben, GMD

Chairman: Bill Briggs, CU-Denver

Chairman: Oliver McBryan, CU-Boulder

Chairman: Charbel Farhat, CU-Boulder


The authors, their affiliation, and titles are listed in the latest
brochure which can be obtained by contacting either program chairman:

Jan Mandel Steve McCormick
(303) 556-4475 (303) 556-4812
(bitnet addresses jmandel@cudenver, smccormick@cudenver will
also work, as well as and

This brochure also contains information and forms for lodging and
registration and other miscellaneous items. (If you have recieved the
earlier form sent last fall, you should soon be getting this latest
one.) PLEASE pay special attention to the lodging information because
the rooms are quick to fill up. If you know others who are going, it
is best to get a deluxe condo that includes two bedrooms with two beds
each and a livingroom and kitchen with a hide-a-bed. They are generally
very nice and most economical, but they go the fastest.


Following is a list of the session times and speakers by last name. In
cases where we were unsure who the speaker will be, we simply list the
first one. Please consult the brochure for other authors and titles.
Updates to the brochure are:
1) G.Lonsdale will speak instead of J.Linden with same title
2) J.Swisshelm, G.Johnson talk cancelled
3) U.Ascher title: 'Semiconductor device simulation'



1:30pm-5:30pm Basic Tutorial: Bill Briggs

7:00pm-8:30pm Software Tutorial: John Adams

8:30pm-10:00pm Reception


8am-noon Flueck Boergers
Lin Cai
McBryan Goertzel
Pasciak Gupta
Quinlan Lee
Szeliski Lubrecht
Tylavsky Mittelmann

8:00pm-9:30pm Advanced Workshop: Achi Brandt


8am-noon Barthelson Adams
Chan Ascher
Chen Bank
Douglas Carey
El-Giar Chin
Farhat Dendy
Pan Ruge

8:00pm-9:30pm Parallel Panel


8am-noon Arakawa Goldstein
Dick Heroux
Dimitriadis Holter
Gustafsson Hoppe
Haenel Khalil
Holland Kornhuber
Koren Mandel

7:30pm-9:30pm Banquet


8am-noon Lallemand Brenner
Linden Kamowitz
Liu Huang
Mavriplis Parter
Mulder Reusken
Napolitano Van der Wees
Wesseling Widlund

5:00pm-10:00pm Special Workshops:
rigorous mode analysis (A.Brandt)
nonlinear multigrid (A.Reusken)
meteorology and oceanography


8am-noon Xu Brandt
McKay Decker
Ong Dervieux
Ruede Lemke
Vanderwalle Padmanabhan
Wittum Witsch
Yserentant Zhu


From: Vincent Broman <>
Date: Mon, 6 Feb 89 12:38:51 pst
Subject: Euler's Constant

Does anyone have Euler's constant, gamma, to at least 35 places,
with a reference for the source? A note about the method
of computation would also be appreciated, if available.


Vincent Broman, code 632, Naval Ocean Systems Center, San Diego, CA 92152, USA
Phone: +1 619 553 1641 Internet: Uucp: sdcsvax!nosc!broman


From: Gary Johnson <>
Date: Mon, 6 Feb 89 20:01:59 EST
Subject: Positions at the North Carolina Supercomputing Center



The North Carolina Supercomputing Center (NCSC) is a state-supported
Supercomputing Center and Research Institute currently under construction
in the Research Triangle Park, North Carolina. NCSC will promote
educational and research activities in high-performance computing
and its applications in science, engineering and other disciplines.
The Center will work in close collaboration with universities and
industry to enhance the economic development of North Carolina and
to support the national high-performance computing initiative.

NCSC's computational environment will include a Cray Y-MP which will
be operational beginning in September 1989. The Center will be accessible
at high bandwidth (T1/T3) by means of North Carolina's existing data
communications network and will be connected to the appropriate national

Qualified applicants with backgrounds in supercomputing operations,
support, engineering, technology or applications research are encouraged
to apply for employment immediately, since staffing plans are currently
being developed. A competitive salary and benefits package will be
offered to the staff of NCSC.

Please send a resume and cover letter to:

North Carolina Supercomputing Center

Human Resources

PO Box 12732

Research Triangle Park, NC 27709-2732

NCSC is a division of the Microelectronics Center of North Carolina (MCNC),
a private, non-profit organization.

MCNC is an Equal Opportunity/Affirmative Action Employer M/F/H/V


From: Giuseppe Paruolo <AGG0%ICINECA.BITNET@Forsythe.Stanford.EDU>
Date: Tue, 07 Feb 89 14:05:12 ITA
Subject: Numerical Algorithms Papers Database

Does anybody know if there exist an "ad-hoc" bibliographical database
containing information on all the published papers in the field of
numerical algorithms for vector and/or parallel computers (and related)?
I already know Inspec but, even if it contains a good amount of
information, the fields covered are too many and its use is quite
Thank you in advance. Giuseppe Paruolo - CINECA
( agg0@icineca.bitnet )


Date: Wed, 8 Feb 89 09:47 EST
Subject: SIAM Conference on Parallel Processing

Conference Date: December 11-13, 1989

Title: Fourth SIAM Conference on Parallel Processing
for Scientific Computing

Organizer: Jack J. Dongarra, Argonne National Laboratory

Abstract Deadline: June 1, 1989

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

Place: Chicago, Illinois

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

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


From: George Corliss <6591CORLISSG%MUCSD.BITNET@Forsythe.Stanford.EDU>
Date: Thu, 9 Feb 89 08:31 CST
Subject: Fortran & Pascal for Scientific Computation

This is an advertisement for Fortran-SC and Pascal-SC. I am not
associated with the authors or distributors. I am a satisfied user;
if you try either of these products, you might like them too.

FORTRAN-SC is an extension of Fortran written at the University of
Karlsruhe for scientific computation. It is particularly suited for
the development of numerical algorithms which deliver high accuracy
and automatically verified results. The language allows the
declaration of functions with arbitrary result type, operator
overloading and definition, and dynamic arrays. Intervals are a
built-in data type. Fortran-SC runs on IBM 370 architecture
machines and uses the IBM program product ACRITH. Fortran-SC
functions as a pre-processor which makes it MUCH easier to program
ACRITH calls. Fortran-SC is available a low cost to ACRITH license
holders. Contact

Dr. Juergen Middel
IBM Deutschland, GmbH
AS Marketing Vektor Systeme
Postfach 80 08 80
D - 7000 Stuttgart 80
West Germany

PASCAL-SC is an extension of Pascal also written at the University
of Karlsruhe for scientific computation. It uses 12 digit decimal
floating point arithmetic and provides optimal arithmetic for many
higher numerical data types such as complex numbers, intervals, and
vectors or matrices over these spaces. Pascal-SC supports
user-defined operator overloading, giving it an Ada-like flavor. It
includes applications packages for linear systems, eigenvalues and
eigenvectors, evaluation and rootfinding for polynomials, and
accurate evaluation of rational expressions. Pascal-SC is available
for about $100 for MS-DOS machines and the Atari-ST. Contact
Dr. Louis B. Rall
5101 Odana Road
Madison, WI 53711
(608) 273 3702

I find both of these tools useful. You might too. Direct questions
or comments to
Dr. George F. Corliss
Dept. Math., Stat. & Comp. Sci
Marquette Univerity
Milwaukee, WI 53233
(414) 224 6599
6591corl@mucsd.BITNET . . . !uwvax!marque!georgec


From: Geoff Chesshire <>
Date: Thu, 9 Feb 89 10:55:09 pst
Subject: Parallel Computational Fluid Dynamics

Implementations and Results Using MIMD Computers
Los Angeles, May 8-9, 1989

ONR, RIACS and INTEL are pleased to participate as sponsors for a conference
designed to discuss recent developments in the use of MIMD parallel computers
in high performance computational fluid dynamics. Each of the invited
speakers has experience and specific results to report using an Intel iPSC/2
system. The issues discussed, however, will pertain to a broad class of
similar MIMD systems.


Antony Jameson
James S McDonnell Distinguished Professor of Aerospace Engineering
Princeton University
"Transonic and Supersonic Flow Calculations Using the Euler Equations"

Anthony T. Patera
Associate Professor, Department of Mechanical Engineering,
Massachusetts Institute of Technology
"Solution of the Incompressible Navier-Stokes Equations on Medium Grained
Parallel Processors"

William T. Thompkins
Manager, Computer Science Laboratory, Northrop Research and Technology Center
"Data Storage Concurrency and Portability: An Object Oriented Approach to
Fluid Mechanics"

Akin Ecer
Professor of Mechanical Engineering, Purdue University at Indianapolis
"Applications of a Hypercube for Solving Three Dimensional Flow Problems"

Mark E. Braaten
Mechanical Engineer, General Electric Research and Development Center
"Solution of Fluid Flows by a Pressure Correction Method on a Hypercube

Geoff Chesshire
Senior Scientist, Applications Research, Intel Scientific Computers
"Fluids Codes and Distributed Memory: Techniques, Tools and Examples"

PANEL DISCUSSION: Are Highly Parallel Systems Ready for Prime Time?

The NASA sponsored panel session will discuss issues in parallel computations
related to a broad class of alternative architectures. Moderator: Horst
Simon, NASA Ames Research Center, Panelists: Creon Levit, NASA Ames Research
Center; Gary Montry, Sandia National Laboratories; Ken Neves, Boeing Computer
Services; Anthony Patera, Massachusetts Institute of Technology; Justin
Rattner, Intel Scientific Computers.


The Parallel CFD organizing committee is soliciting contributed presentations
of 10 minutes or less on subjects which will complement the in-depth titled
papers described above. Proposed title and abstract (not to exceed 250 words)
should be sent to Lynn McDaniel, AlTek/Parallel CFD, 8321 N. Woolsey,
Portland, Oregon 97203. Abstracts deadline February 28, 1989. Notification
of acceptance March 13, 1989.


Radisson Plaza Hotel
1400 Park View Avenue
Manhattan Beach, California 90266
(3 miles south of LAX Airport, just west off I-405 at Rosecrans)

8:30 a.m. to 6:00 p.m. Monday, May 8
8:30 a.m. to 12:00 p.m. Tuesday, May 9

A block of rooms has been reserved at the Radisson Plaza Hotel. Please
contact the hotel at 1-800-228-9822 to make hotel reservations. Be sure to
identify yourself as a participant in Parallel CFD to receive the discounted
conference rate.


Registration for Parallel CFD is limited to 200 people. Advance registrations
of $100 will be accepted until March 10, 1989. Registration at the
conference, $125 will be on a first come, first served basis for those without
reservations. Day of Conference registration will begin at 7:00 a.m. May 8,
1989. If you have questions regarding the conference, please call Lynn
McDaniel at (503) 289-7735.


From: Gene H. Golub <>
Date: Sat, 11 Feb 1989 17:17:39 PST
Subject: Wilkinson - SIAM Prize

The 1989 Wilkinson Prize in Numerical Analysis and Scientific
Computing is being awarded by SIAM to Paul Van Dooren of Philips in
Brussels. Paul's main contributions have been to the development and
analysis of algorithms for linear problems in systems theory. This is
a very vital and active area of research at this time. Wilkinson had
a high opinion of Paul; he was on his the reading committee and often
said only a few people realized the quality of this work. Paul was
subsequently awarded the Householder Prize.

Beyond Paul's research abilities is his lovely sense of humor and human
qualities. Congratulations, Paul!

The award will be given at the national SIAM meeting in San Diego.
Hope to see you there,

Gene Golub


From: Gene H. Golub <>
Date: Sat, 11 Feb 1989 17:20:33 PST
Subject: FAX

Have you tried the FAX technology? I was in Europe a few months ago and
I needed by notes which were in Australia. The next day my own hand written
notes appeared before me and I was able to give my lecture!

I'm trying to collect FAX addresses and I have a few. If you have one, please
send it to me.


Gene Golub
Director, Scientific Computing/ Computational Math
Office: Margaret Jacks Hall, Room 306
Office Phone: 415/723-3124 Home Phone: 415/323-0105
FAX number: (415) 725-7411


From: Gene H. Golub <>
Date: Sat, 11 Feb 1989 17:39:16 PST
Subject: Kudos to Bill Gear

Here's kudos to Bill Gear who completed the job of President of SIAM in
December. (He's now officially Past-President.) In case you didn't notice,
SIAM seems to be as thriving as ever and much of that is due to Bill's
efforts. As his predecessor, I sat back with admiration in seeing him handle
the many complexities of the job with ease. Here's a quote from Margaret
Wright who is on the Council and observed Bill first hand.

"In addition to Bill's overall vision for and leadership of SIAM, his
remarkable skills in ``chairmanship'' have been evident at meetings of the
SIAM Council and Board. An a posteriori analysis of meetings chaired by Bill
reveals that he always manages to achieve stable and accurate
solutions, even to problems that might appear at first to be ill-posed!
The meetings are well organized, smoothly run and productive---never
meandering or regimented. Bill's equanimity and sense of humor
maintain a balanced and reasonable atmosphere, even in the midst of
disagreements among the participants."

Bill not only was Prez of SIAM but he was the chairman of his
department and managing editor of SISSC. (How did he do it?)

Thanks, Bill for all your efforts. We have all benefited.
I hope you carry on all your activities and research as successfully
as you have in the past.



End of NA Digest