Remarks on Supercomputing in Germany

1 ABSTRACT Simulation has become one pillar of scientific research in the last years. Much as experimental work has changed scientific thinking when it complemented reasoning and logical thought, simulation has started to change the way science works. And just like experiments have not replaced but complemented logic, simulation will not replace any of the other two methods but will complement them. There are a number of problems that come with this. Many simulations can not be verified by experiments and thus have to be taken at face value. Furthermore working in the field of simulation we are aware of the limitations of our physical and mathematical models and know that we have to improve them in order to get better and more reliable simulation results. There is one thing, however, that has not changed over the last years: Simulation requires compute power and the increased complexity of simulations requires an increased level of performance. Supercomputing – or High Performance Computing (HPC) – is therefore a technology that is at the forefront of development in simulation research. Because of its nature as a leading edge technology it not only serves as an enabler providing the necessary compute power for mathematicians and application experts. Supercomputing also has the role of a path finder in the field. To some extent it can be compared to space technology or the formula 1 – collateral scientific developments of which have consistently found their way into our every day world. Given the crucial importance of supercomputing as a spearhead of science this article sets out to look at the role of supercomputing in Germany and the German research community. For this we will lay out the current situation in supercomputing and aim to give a comparison with other industrially developed countries. We will also have to consider the future challenges and chances in supercomputing in a larger Europe that has to compete with the USA and Asia.