\documentstyle[twoside,epsf]{article}

\pagestyle{empty}

\newfont{\fit}{cmti10} % Text wg. baselineskip machen wir das so !
% ---------  Definitionen und Abkuerzungen  ----------------
\def\topfive{{\sc Top500}} % Small Caps !!
\def\topone{{\sc Top100}} % Small Caps !!
\def\toptwf{{\sc Top25}} % Small Caps !!
\def\linpack{{\sc Linpack}} % Small Caps !!
\def\rmax{$\fit R_{max}$}
\def\rpeak{$\fit R_{peak}$}
\def\nhalf{$\fit N_{1/2}$}
\def\nmax{$\fit N_ {max}$}
\def\flops{flop/s}

\begin{document}
\title{Preface}
\author{}
\maketitle

\section*{The Aims}
This report is a snapshot of the state of supercomputer 
installations in the world. 
It is  based on the \topfive\ list that was published in 
November 1995 and includes trends from  
the previous lists from June 1993 till November 1995.

Statistics on high-performance computers are of major 
interest to manufacturers, users,
potential users, and decision makers in universities, 
government, and industry.  
These people wish to know not only the number of systems installed, 
but also the locations of the various supercomputers within the 
high-performance computing community, and the applications for which  
a computer system is used. 
Such statistics provide a better understanding of the  
high-performance market and can facilitate the exchange of data and software. 


In the past, various system counts of the 
major vector computer manufacturers  by  continents and countries have 
been published. 
Such records have several limitations, however.  
The data was  difficult to obtain, and often was not reliable. 
Most important, more extensive  
statistics (than simply a list of manufacturers' names) are now required 
because of the  diversification of  
supercomputers, the enormous performance difference 
between low-end and high-end models, the  
increasing availability of massively parallel processing (MPP) 
systems, and the strong  increase in computing power of the 
high-end models of workstation suppliers including symmetric multiprocessing
(SMP)  
systems.  


This report is meant as an interface between the \topfive\ 
list and the reader who wishes more  
background information and explanation. 
Here various experts present detailed analyses of the  
\topfive\ and discuss the changes that have occurred in the 
supercomputing market over the  past year.

We plan to continue to update this report annually and to distribute it 
widely to the high-performance  computing community. 
The first \topfive\ Reports were covering the situation in 1993 and 1994.

\section*{The Contents}

This report consists of ten articles.
The first five articles present a detailed analysis of the 
high-performance computing  
situation as of November 1995.
The next four articles focus on different aspects 
of the hardware and software  
of the systems, which can be seen in the \topfive\ . 
The final article is a complete  
reprint of the November 1995 issues of the ``\topfive\ Supercomputer Sites,''
which provides the basis of this report. 


Meuer and Strohmaier analyze in the first article the 
general worldwide trends, which are  revealed by the 
six releases of the \topfive\ published in the past three years. 
They present  the changes over time with respect to geography, 
manufacturers, applications, architectures,  
and technology. 
%
Dongarra and Simon present an in-depth analysis of the 
U.S.~situation of the field of
high-performance computing.
%
Hoffman and Schnepf give an overview of the Japanese installations and
Japanese  vendors and the differences to the overall market. 
%
Harms discusses the European situation and provide a brief 
summary on computing in the 
United Kingdom, France, Germany, and the Benelux nations.
%
Bez and Simon present a short description of the 25 centers
with the highest accumulated performance installed.
%
Van der Steen summarizes the new architectures of the 
different systems in the \topfive\, giving a  concise description for 
each architecture. 
%
Dongarra and Walker present an overview of the  
new MPI standard, which is at present  the 
most promising attempts to create a   
standardized programming environement 
for parallel computers with distributed memory. 
%
Nagel et al.~descibe in their article 
a new graphical interface for tracing, 
debugging and performance tuning for MPI. 
%
As a case study for Cluster computing  
Mierendorf, Sch\"uller and Trottenberg
present their results for running the 
IFS weather prediction model of the ECMWF 
on a cluster of four full blown C90s.
For this work they were awarded the {SuParCup'95} 
at the {\it Mannheim Supercomputer Seminar '95}.


\section*{The Audience}

The report has been prepared for the high-performance computing 
community in general, and  specifically for managers of 
supercomputer centers, users of supercomputers,  
computer/supercomputer manufacturers, consultants, professional 
market analysts, decision  makers, politicians, Wall Street analysts, 
computer science people, and students.

\section*{Acknowledgments}

Without the help of high-performance experts, 
computational scientists, manufacturers, and the  
Internet community, our \topfive---the basis of this report---could 
not have been compiled. 
We  cordially thank all colleagues supporting us, and 
we ask for their continued support in order  
to present  future \topfive\ lists and reports like the one presented here.

Mannheim / Tennessee, November 17, 1995

\makebox[\textwidth]{Jack J.~Dongarra \hfill$\bullet$\hfill 
Hans-Werner Meuer \hfill$\bullet$\hfill Erich Strohmaier}       


\end{document}
