Stuttgart Research Centre for Simulation Technology and Cluster of Excellence in Simulation Technology
From Isolated Numerical Approaches to an Integrative Systems Science
In the 21st century´s society, simulation
technology has become completely
indispensable, it dominates all areas of
life. The cluster of Excellence in Simulation
Technology, approved within the
German national "Initiative of Excellence"
is embedded in the Stuttgart Research
Centre for Simulation Technology. It represents
a massive platform for further
developing scientific methods and techniques
in all branches of modeling and
simulation techniques. HLRS is taking a
lead in the cluster being responsible for
High Performance Computing research.
Scientific Goals
Currently, Simulation Technology is one
of the most challenging research areas
in contemporary engineering and natural
sciences. It offers solutions for a broad
variety of applications with an enormous
economic and scientific impact. Recent
and future developments in technology
and society pose challenges which require
Simulation Technology to advance
to a new level. The Cluster of Excellence
in Simulation Technology will strengthen
the sustainable evolution of research
areas that are of fundamental importance
in both science and economy, as
well as to conceive new and visionary
applications of simulation sciences. The
goal of SimTech is to further intensify the
university´s engagement in the field of
Simulation Technology in order to maintain
and expand the excellent standing in
this important field in a dedicated intersectorial
structure. Establishing integrative
prediction methods and solutions
for future challenges, the cluster concentrates
on six methodical Research
Areas ranging from Molecular Dynamics
and Advanced Mechanics via Computational
Mathematics and Systems Analysis
to Data Management and Interactive
Visualisation as well as to Hybrid High-
Performance Computing Systems and
Simulation Software Engineering. An
Integrative Platform acts as a bracket of
reflection and evaluation. Through transfer
of scientific results to applications,
the cluster will furthermore contribute
to the economic benefit of our society.
 |
| Figure 1: Simulation
of a β-lactamase
enzyme in water |
The Role of HLRS in SimTech
The director of HLRS Prof. Michael
Resch is a Principal Investigator and
one of the two Coordinators responsible
for the Research Area: Hybrid High Performance Computing. The focus of
the work of HLRS in the cluster of excellence
is on Parallel Programming Models
and Tools. Models and tools to handle
large scale hybrid systems are needed
to harvest the potential of millions of
components in future supercomputing
systems hiding the complexity from the
non-expert user. HLRS has therefore
set the following short and medium term
goals for the work in the cluster.
- Development of parallel programming
models capable of extending existing
approaches to large scale systems
and later to hybrid systems
- Development of methods and tools to
support large scale parallel programming
and further to hybrid systems
- Development of resource
management methods to describe
and manage workflows for large
scale simulations on distributed and
hybrid systems
 |
| Figure 2: Temperature
distribution [°C]
as 3D projection with
1,300° C isosurface
(grey) to visualize the
fire location |
Resource Management
The efficient utilization of concepts of
parallelization confronts the scientific
software developer with many problems
when having to solve complex applications
on distributed heterogeneous resources
or Grid resources. Very often
this is due to the fact that the subject
areas are already highly complex in
themselves. So far for many fields there
has been no systematic approach for
the development of complex time-critical
Grid applications. HLRS will be devoted
- to the study of the peculiarities of
the organization of distributed heterogeneous
computation environments
and of their components in order to
solve the problem of the efficient
execution of complex applications
- to the development of resource
management methods guaranteeing
a certain level of functional reliability
for time-critical applications (applications
that have to be done within a
certain time like business simulations)
and real-time simulations (applications
that are part of a process and
run permanently like weather forecast)
to improve the quality of the
handling, the design, the deployment
and the execution of complex largescale
simulations
HLRS already developed a basic concept
for the solution of complex applications
in scientific and engineering experiments
on distributed heterogeneous
resources. Part of this work was done
in the European research projects Grid-
Coord and CoreGrid. The concept comprises
two main components: a parallel
language GriCoL for the description of
complex scientific experiments and the
Science Experimental Grid Laboratory
(SEGL) for the control and execution of
complex experiments.
References
• Natalia Currle-Linde
• Michael Resch
University of Stuttgart, HLRS
 top
|
index
prev