Abstract Dr. Klaus Ecker

Why do we need high performance computation?

Applications such as global warming prediction, seismic and volcanic activity prediction, food and health biology, or economic globalization are well-known problems we are confronted with almost every day by the media. Some people may be aware of the struggle behind the attempt to find solutions to these problems. But only experts know of the complexity of models for these applications and the huge computational expenses required for simulating the applications and finding good solutions. Mathematically accurate models must be defined that can be used as a basis for simulating the particular application; simulation algorithms and fast implementations are to be developed. Technically, proper computer architectures for performing the simulations in acceptable time, such as grids or clouds, and their operational organization need to be provided.
The application area of our particular interest is computational biology, where genome and protein research is one of the prevailing activities. The rapid developments enable progresses in human medicine unthinkable a few years ago, help to tackle the food problem, and many more. But these progresses do not appear from nowhere. Considerable part of the development went hand in hand with the availability of increasingly powerful computers that enabled researchers to manage soaring amounts of data and process them in increasingly complex computations. Genome sequencing, structural genome analysis, protein structure prediction, organization of protein databases, and phylogenic predictions are just a few of the many problems whose solutions contribute to the successes in modern biology.
In this presentation we will justify the needs for high performance computer power in computational biology and address technical concepts for its realization.