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 2, Number 7 (July 29, 1992)

Today's topics:

     Question about multigrid for unstructured grids
     Summary of workshop on C++
     Seventh Conference on Domain Decomposition Methods
     Paper entitled Kinematics of Multigrid Monte Carlo
     MADPACK Status (versions 2 and 4 available)
     Incoming directory status


Date: Fri, 19 Jun 92 10:10:08 -0700
From: (Horst D. Simon)
Subject: question for mgnet readers

I would like to find out what ideas have been proposed or implemented
in the area of multigrid for unstructured grids. In particular I would
like to obtain references about algorithms for coarsening existing
unstructured grids. But any general references about unstructured
multigrid would be useful as well.

Horst Simon

    Editor's Note:  I think many readers would like to know this, too.
    -------------   Please cc this list.


From: Ulrich Ruede 
Subject: Summary of workshop on C++

On April 14th, at the Copper Mountain SIAM conference on iterative
methods, a workshop was held to present C++ as a useful language for
Scientific Computing.  The workshop was organized by Dan Quinlan and
included a panel of seven participants: Dinshaw Balsara, from Johns
Hopkins University; Doug Clarkson, president of Dyad Software
Corporation (developers of the M++ array class library); Carl
Kesselman, from Caltech; James S. Peery, from Sandia National
Laboratory; Daniel Quinlan, from University of Colorado at Denver;
Allen C. Robinson, from Sandia National Laboratory; and Ulrich Ruede,
from Technische Universitaet Muenchen.

   C++ as a language for scientific computing has recently been gaining
increased attention. Besides providing the programming concepts of an
advanced language it can be highly efficient, is widely available and
is very portable. Its features make it almost ideally suited for the
tasks in mathematical and scientific software development.  The class
concept and operator overloading make the language extensible so that
user defined data structures can be handled the same way as built-in
data types. Thus all the FORTRAN-90 array language functionality can be
provided on the library level. Similarly, most standardization problems
for numerical libraries in FORTRAN would be much easier in C++, because
the abstraction mechanisms of C++ allow more natural library interfaces
that are not obscured by work-arounds for language deficiencies.

A brief summary of the workshop is available by ftp from the mgnet server.

Ulrich Ruede


From: Jinchao Xu 
Date: Tue, 28 Jul 92 16:17:22 -0400
Subject: Seventh Conference on Domain Decomposition Methods

The Seventh International Conference on Domain Decomposition Methods will be
held at Penn State during the four day period October 27-30, 1993.  If you
want to receive further information, please contact:  R. Manning (secretary),
Dept. of Math., Penn State Univ., University Park, PA 16802. Tel: 814-865-7527;
Fax: 814-865-3735; Email:


From: Martin Grabenstein 
Date:    Wed, 29 Jul 92 10:09:29 MESZ
Subject: Paper entitled Kinematics of Multigrid Monte Carlo

Kinematics of Multigrid Monte Carlo,                                            
M.Grabenstein and K.Pinn, 28 pages, 8 ps-figures, preprint DESY 92-094          
We study the kinematics of multigrid Monte Carlo algorithms by means of         
acceptance rates for nonlocal Metropolis update proposals.                      
An approximation formula for acceptance rates is derived.  We present           
a comparison of different coarse-to-fine interpolation schemes in free          
field theory, where the formula is exact.                                       
The predictions of the approximation formula for several interacting            
models are well confirmed by Monte Carlo simulations. The following             
rule is found: For a critical model with fundamental Hamiltonian                
H(phi), absence of critical slowing down can only be expected if the            
expansion of  in terms of the shift psi contains no               
relevant (mass) term.                                                           
We also introduce a multigrid update procedure for                              
nonabelian lattice gauge theory and study the acceptance rates for              
gauge group SU(2) in four dimensions.                                           

    Editor's Note:  Installed in the papers/Grabenstein-Pinn directory.


From: Craig Douglas 
Subject: MADPACK Status (versions 2 and 4 available)

I have moved version 2 of madpack (the public domain version) to the directory
madpack2 and have installed a madpack4 directory with the recently released
IBM copyrighted codes damg and dpmg.  There is a short questionaire associated
with version 4 that I hope people will fill out even if they decide not use
the codes.  This should be returned to me at Yale, not IBM.  This is for my
benefit.  If there is enough interest, I will put a summary in the madpack4

Note that the documentation files are not present in version 4.  There are
several thousand lines of comments at the head of both programs that should
help.  Anyone serious about using these codes should contact me for
preliminary LaTeX files that provide a little more insight.  (I did not want
to waste my time writing papers if the codes were not going to be released to
the public in source form.)


From: Craig Douglas 
Subject: Incoming directory status

I have opened the mgnet/incoming directory again.  I will have to shut it down
whenever I am out of town.  So, please start putting objects there again,
except during August 1-7...

While I am on sabbatical full time at Yale this year, things should work
better than they have the past two months.  My apologies to those who have had
difficulties this month in particular.


End of MGNet Digest