Send mail to:             for the digests or bakeoff
            for comments or help
 Current editor:  Craig Douglas       
Anonymous ftp repository: (

Today's editor:  Craig Douglas (

Volume 3, Number 11 (November 30, 1993)

Today's topics:

     Hackbusch's Iterative Methods Book
     New Multigrid Book (Bramble)
     New Multigrid Book (Ruede)
     2 MGNet Paper Submissions
     Sixth Copper Mountain Multigrid Proceedings
     Breckinridge '94, Copper Mountain '95, Tahiti?
     The bibliography database on MGNet


Date: Thu, 4 Nov 93 10:29:42 PST
From: (Herb Keller)

   The NSF Science and Technology Center for Research on
   Parallel Computation (CRPC) presents a one-day short course
   on "Parallel Computation: Practice, Perspectives and Potential",
   followed by demos and tour of parallel processing facilities,
   January 24, 1994, in Ramo Auditorium at the California Institute
   of Technology, Pasadena, CA.  Speakers:  K.M. Chandy, J. Dennis,
   J. Dongarra, G.C. Fox, K. Kennedy, D. Meiron, P. Messina.  For more
   information contact JoAnn Boyd, email:,
   telephone: 818-395-4562, Fax: 818-683-3549.


Date: Fri, 5 Nov 93 18:13:22 EST
From: (To The Many Nice People)
Subject: Hackbusch's Iterative Methods Book

Many people responded to Tony Chan's request in the last issue for information
on how to order Wolfgang Hackbusch's recently published in English book on
iterative methods.  Thank you all very much.

Here is the entry in the MGNet bibliography:

  author =      "W. Hackbusch",
  title =       "Iterative Solution of Large Sparse Systems of Equations",
  publisher =   "Springer--Verlag",
  address =     "Berlin",
  year =        "1993"

It has 380 pages, 11 illustrations, and 40 tables.
ISBN 0-387-54795-9 (hardcover).

In North America, it can be ordered from Springer using the following:

    telephone:  800-springer  or in New Jersey  201-348-4033
    fax:        201-348-4505

Springer advertised this book on page 5 of the August, 1993 SIAM News.


Date: Mon, 15 Nov 93 11:02:42 EST
From: (James Bramble)
Subject: New Multigrid Book

My notes are now out in book form in series "Pitman research notes
in mathematics". It is Number 294 and is called "Multigrid Methods".
The ISBN number is 0 582 23435 2.

Naturally, there were some screw-ups. Here is an errata sheet for those

\input amstex
\centerline{{\bf Errata for Multigrid Methods}}
\centerline{{\bf by James H. Bramble}}
\medskip\advance\parskip by 3pt

1. The reference numbers in the text from 55 to 130 should be increased by one.

 2. The reference numbers in the text after 130 should be increased by two.

3. The first display on page 111 should read
A_k(u,v) = \sum_{\tau\in\Cal T_{h_k}}\sum_{i,j} \int_\tau
   \frac{\partial^2 u}{\partial x_i\partial x_j} \
          \frac{\partial^2 v}{\partial x_i\partial x_j} \,dx.


The following is some propaganda for multigrid in general and my notes
in particular which is printed on the back cover.

"Every researcher working in the area of the numerical solution of partial 
differential equations, as well as practitioners in the field of scientific 
computations, should have at least some familiarity with this powerful 
technique. Multigrid methods are among the most efficient iterative methods 
for the solution of linear systems which arise in many large scale scientific 
calculations. Consequently, the invention of the multigrid method in the 
1960's and its subsequent development up until the present is an extremely 
important area of research with great practical implications. This book 
presents old and new results concerning the rates of convergence of multigrid 

The first part of these notes contains an abstract multigrid theory and its 
connection with general linear and nonlinear iterative methods. Several 
sections are devoted to important, typical applications.   An emphasis is 
placed on how the latest results can influence  the design of effective, 
robust techniques."


Date:   Tue, 16 Nov 1993 19:22:09 +0100
From: Ulrich Ruede 
Subject: Book announcement

I am pleased to announce the publication of my book:


                Ulrich Ruede
        Frontiers in Applied Mathematics, Vol. 13,
        SIAM, Philadelphia, 1993, ISBN 0-89871-320-X
        140 pages, includes 116 bibliographical references and an index
        List price: $23 (SIAM members: $18.40)

In this monograph I attempt to present a unified approach to the
development of multilevel adaptive methods for the solution of partial
differential equations, including

        * a rigorous mathematical foundation based on results in the theory
          of function spaces and in approximation theory
        * algorithmic aspects including the
          multilevel adaptive relaxation method
        * software develoment issues and data structures using the
          principles of object oriented programming

The book can be ordered directly from SIAM (e-mail:

        Ulrich Ruede
        Institut fuer Informatik
        Technische Universitaet
        D-80290 Muenchen, Germany


        Fachbereich Mathematik
        Technische Universitaet Chemnitz-Zwickau
        PSF 964
        D-09009 Chemnitz, Germany


Date: Fri, 19 Nov 93 10:10:16 +0100
From: Thor Gjesdal 
Subject: 2 MGNet Paper Submissions

                      Programming Multigrid in Fortran90

                                 Thor Gjesdal


We describe a simple one-dimensional multigrid Poisson solver in order to
illustrate some of the new features of Fortran90, and how they can be used in
multigrid methods.

                                   * * * *

                  Analysis of a New Red-Black Ordering for 
               Gauss-Seidel Smoothing in Cell-Centred Multigrid

                                 Thor Gjesdal


Based on numerical experiments with a cell-centred multigrid Poisson solver in
one and two dimensions, we propose to use Alternating Coarse-Line Zebra
ordering as a robust parallel smoother for the 2D case.  The algorithm is
presented based on a heuristic discussion, and the convergence rates obtained
with different smoothers are compared.  Local mode analysis is used to
estimate the smoothing factor and two-level convergence for the method.

Math. Reviews 1991 Subject Classification: 65N22, 65N55, 65Y05

    Editor's Note: in mgnet/papers/Gjesdal/mg_f90.abstract,
    -------------     mgnet/papers/Gjesdal/mg_f90.dvi,
                      mgnet/papers/Gjesdal/new_rb.abstract, and


Date: Tue, 23 Nov 1993 13:12:23 -0500
From: Duane Melson 
Subject: Sixth Copper Mountain Multigrid Proceedings

The proceedings of the Sixth Copper Mountain Conference on Multigrid Methods
is currently being printed by NASA.  Copies will be mailed to the participants
at the address they provided at the conference.  They should arrive by
Christmas.If anyone does not get their copy by the first of the year, please
let me know by e-mail.  I will have some extra copies that I can send out
(only 50).

Duane Melson
NASA, Langley Research Center
(804) 864-2227

    Editor's Note:  Rumor has it that this is 800+ pages in 2 volumes.


Date: Mon, 29 Nov 1993 09:12:12 -0500
From: (Craig Douglas)
Subject: Breckinridge '94, Copper Mountain '95, Tahiti?

There has been some confusion on where the 7th Copper Mountain Multigrid
conference will be held.  It will be at the usual place in Copper Mountain in
April, 1995 (hopefully early with lots of snow).

The deadline for submitting an abstract (1-5 pages) to the Breckinridge
Iterative Methods conference is December 15th.  Send it to


Date: Mon, 29 Nov 1993 09:14:32 -0500
From: (Craig Douglas)
Subject: The bibliography database on MGNet

The bibliography has had numerous corrections made to it so far this month.
One of the most common errors in it was in entries for proceedings of
conferences which were published in a different year than the conference.
Hopefully, these have all been fixed.

In addition, it has grown by several more entries.  All of the entries from
the first 5 domain decomposition symposia are in there along with the relevant
citations from the papers.  The main multigrid conference proceedings are in
the process of having the same done for them.

A few of you are still bashful about putting all of your multigrid and/or
domain decomposition papers into the list.  Just because S. F. McCormick
starts on page 47 does not mean that we cannot make that much further along in
the printed document by next month.  Particlularly if certain people with
initials AB, TFC, JED, REE, PWH, WH, BK, SFM, SVP, JEP, and OBW would
contribute their missing ones (if any)...

Many thanks to those of you who have taken the time to send in entries or


End of MGNet Digest