In case I you are interested in this stuff I would like to ask you to
identify yourself by your e-mail.
This will just a statistic for me to see who is interested in my stuff.
In den letzten zwei Jahren haben sich auf dem Gebiet des verteilten
objektorientierten Programmierens drei verschiedene Objektmodellen
etabliert:
CORBA - Common Object Request Broker Architecture
COM/DCOM - (Distributed) Component Object Model
JavaRMI - Java Remote Method Invocation
Jedes dieser Modelle wird im Moment in relativ klar voneinander
abgegrenzten Bereichen eingesetzt (COM/DCOM ausschliesslich bei
Microsoft, CORBA beim Aufbau von verteilten Client-Server
Systemen unter UNIX, JavaRMI bei der Entwicklung WWW basierter
Anwendungen).
Obwohl die Entwickler dieser Modelle auf unterschiedliche
Herangehensweisen bei der Erstellung verweisen, lassen
sich dennoch eine Reihe Gemeinsamkeiten aufzeigen. In Vorbereitung
eines Vergleichs sollen dazu CORBA, COM/DCOM und JavaRMI im Detail
vorgestellt werden.
Ein Nachteil tritt bei der Einordnung in die "klassischen" Konzepte
des objektorientierten Programmierens hervor: CORBA und COM/DCOM verfügen
nicht über die Möglichkeit der Implementationsvererbung.
Möglichkeiten zur Beseitigung dieses Mankos sollen anhand eines Systems,
welches Konzepte von CORBA, COM/DCOM und JavaRMI in sich zu vereint diskutiert
werden. Fuer dieses System sollen weiterhin Mechanismen zur
Zustandsmodellierung, zum Scheduling und zur Migration von Instanzen
eroertert werden.
The Convex SPP/1000 Exemplar (Scalar Parallel Processing) is a shared memory
parallel architecture supercomputer based on HP PA7100 processors.
Each processor got 1 MByte Cache.
It may work as stand alone (no host is necessary) running HP-UX. It is organized in a set
of maximum 16 hyper-nodes, each containing up to 8 processors and up to
2 GByte shared memory with latency around 2.5 us over the full system.
The hyper-nodes are connected via a CTI (Convex Toroidal Interconnect).
According to the Convex Exemplar Programming Guide I tried out some small
examples and collected the most important function calls to a small number
of sheets.
It is well known that Atomic Force Microscopes (AFM) are sensitive to
sound and vibration. To quantify this sensitivity data were recorded
from an AFM in a controlled sound environment.
Using these data some qualitative and quantitative conclusions are
done. Also some ideas how to prevent an AFM from such distortions are
discussed.
This was my final project at the Technical University of Dresden.
In this publication the mathematical behavior of the most important image processing
functions are discussed. A lot of investigation are done on the Discrete Fourier Transformation
which is a very powerful tool for image processing. It was possible to give some error
estimation for the DFT. Furthermore filter functions are discussed both in time and frequency
domain. Some new filter functions are developed which proof the structure of an image
to be the same before and after transformation.
Image processing is discussed also as part of system theory.
A polynomial fit algorithm is introduced using a QR factorization for solving. The problems
of drift and creep are also discussed and a model is given (differing a little bit from that
of J.F. Joergensen and not proved by experiments).
A great part deals with the visualization of the measured data. Therefore several interpolation
algorithms are tested (Bi-linear and Tensor spline interpolation).
Finally the XRTMA program is introduced, which is an image processing tool based on the
things discovered in the previous sections. It uses a X11/OSF-Motif Interface and runs on
IBM/PC under Linux, IBM RS/6000 AIX and Convex C-series.
The program is capable to deal with IFW- (Institut fuer Festkoerperphysik und Werkstofforschung
Dresden e.V.), NanoScope III-, plain text- and DFM- (Danisch Institute of Fundamental Metrology)
formatted data. Different STM scanning modes (CCM, CHM) and AFM modes can be analyzed.
The program also has the opportunity to handle potentiometric and spectroscopic data.
A great advantage is the scalability for special problems. The program comes with a very strict
distinction between user interface and functionality. By using the included library it is
possible to adapt the program to special problems. This is also possible by using an integrated
script language.
diploma, XRTMA Documentation,
compressed PostScript 324 kByte (in German) the XRTMA program,
compressed archive 2.5 MByte,
pre-compiled for Linux, AIX and Convex-OS, including sources
Note: The software is restricted by the rules for diploma works at the Technical University
of Dresden and the Institute for Solid Physics and Surface Science IFW Dresden.
Copying and using the stuff is highly welcome but not selling it to a third party.
In this paper the illumination models of Lampert and Phong are described,
The ray tracing principle is discussed and a description of the implementation
in the "spm" image processing tool of J. F. Joergensen (my supervisor at Danish
Institute of Fundamental Metrology, Lyngby, Denmark) is given.
When starting X and OSF/Motif programming during my time in Denmark I where punished
several times by using constants which I not knew nor could find somewhere (and believe me
I had a look at nearly each page of the 1.5 meters documentation from IBM for the AIX
system). This forced me to compile often used function with often used constants to a
simple calling interface. Also to make it easy to port graphics code between X and non-X
applications I wrote a small graphics module which uses X intrinsic calls in combination
with the Motif toolkit.
This report was given at the Research Center
Rossendorf as seminar at the Department for Communication and Data Processing (FVTK).
Here the basic principles of OOP are described, including a short history, the main concepts,
the principle of inheritance, polymorphism and attribute/method scope rules.
In the second part a short introduction to C++ is given, including function overloading,
class declaration, inheritance, virtual and abstract classes, generating objects, friends
and operator overloading, some simple examples and some tasks for learning C++.
In summer term 1997 we held an introduction seminar to object oriented programming. The seminar
covers common aspects of the OO paradigm like abstract data types, polymorphism and inheritance.
We introduce several OO languages, e.g. Oberon, Java, C++, Smalltalk, Turbo-Pascal, Ada.
We also recommend some small problems to be solved in an OO style.
In this talk a short introduction to what VR means is given. The talk covers aspects of
influencing human senses, available systems and software as. In the second part a bases for
creating distributed virtual environments is discussed and two different systems are introduced
(DIS from U.S. Navy and my own). Before a demonstration of a working DVRE (distributed virtual
reality environment) is taking place, the VRML language is introduced.
Nils Schmeisser, Dr. Manfred Kunicke, Dr. Karl-Heinz Heinig
The report was accepted by the High Performance Computing Users Group Conference '98.
The continuing exponential increase in computer power together with the
recent developments of very efficient numerical procedures allow nowadays
to perform predictive atomic-scale computer simulations for material science.
This holds especially for advanced microelectronic devices where functional
units consists more and more often of 10^6 atoms or even less. In this
situation the design of new materials and devices is more and more frequently
supported by atomic-scale computer simulations.
A kinetic 3D Monte-Carlo code based on stochastic probabilistic two-centre
cellular automaton using a double bookkeeping technique, one in the particle
vector and the other in the lattice space was originally coded in PASCAL and
tested on an INTEL PC, later in C and on HP workstations.
This implementation should be speeded up considerably in order to undertake
real scientific simulations. The only way to get the needed speedup is
parallelisation.
We started with a look at the given sequential code and run it on our
S-Class Server. While optimising we had to understand a lot of things, e.g.
the double bookkeeping technique, so that we could do first steps of scalar
optimisation during this process.
The most simple idea of a parallel approach is distributing the lattice across
the processors. But because of the physical problem modelling the growth of
clusters of implanted particles in the lattice such an approach will lead to
load imbalances and destroy the effect of parallelisation.
With an another approach we tried to distribute the work done in the particle
space instead of the lattice space. We developed a graph-theoretical approach
based on skeletons to find out a work distribution across 2^n processors that
assures a good load-balance. The disadvantage of this approach is the work
needed to compute the distribution which is of order O(n^3) compared to the
simulation which is of order O(n).
At this point it turned out the given model was not suitable for
parallelising the algorithm. So we had feed-back with the physicist and
together we developed a model of the physical process where the algorithm
which solves the problem could be formulated in parallel. The new model
allowed us to divide the solution into three main steps, where two of them can
be done parallel and will therefor improve the speed of the algorithm.
We implemented this algorithm in a mixed language modularised program using
FORTRAN for the computational part and C for the I/O part. The program runs
on our S-Class server using maximum 8 processors.
In our presentation the methodology of creating an optimised parallel code for
the Monte-Carlo simulation of Ostwald Ripening, the schema of the
parallel code and first real physical simulation results will be shown.
This small tool can display text based data like gnuplot. It is possible to choose 1D and 2D
diagram style with color overloading. Single points can be checked out by clicking.
It looks like who it is who. The program displays all currently logged in users (by
examining the files in /etc. You can get information on each user and on the whole
system.
This small tools displays a molecule lattices in 3D. The lattice can be rotated around
all axis. The Woboo Electronics
LCD shutter glasses are supported for real 3D display.
The driver is included to the archive.
This X11 front-end to pgp includes an integrated environment with an editor and controls
to pgp as well as to the mail program.
It requires pgp and the UNIX command line mail to be installed.
In this paper we discuss several methods displaying a surface given in an explizit form
of f:RxR->R:f(x,y)=z. We start with different types of projection (looking backward to
the time we wrote it we have to state that now we would use homogenous coordinates instead
of normal 3D vector space). Then we demonstrate different visual types (wireframe, iso-lines,
solid surfaces). Finally we describe some methods for visibility checking (hidden-line,
depth sort and an analytical approach). We also added information on how to make a real 3D
image by the use of analglyph (red/blue) glasses. The paper comes together with an unit for
Turbo-Pascal.