The most scientifically important and economically promising research frontiers in the 21st century will be conquered by those most skilled with advanced computing technologies and computational science applications.
This statement, one of the key conclusions of the 2005 PITAC report in the USA, depicts the field of research to be addressed by the Munich Centre of Advanced Computing (MAC) as well as its high scientific and economic relevance. Computational Sciences or Computational Science and Engineering (CSE) have become a highly interdisciplinary, challenging, and thriving scientific domain of increasing importance and visibility in research, development, and education. Their essence primarily involves modelling and computations – in particular (numerical) simulations – on computer systems ranging from standard PC to supercomputers. Thus, CSE applications have always been a driving force of computing technologies, especially progressing the field of supercomputing.
However, at present, an increasing gap is observed world-wide between what should be possible in theory due to recent advancements in algorithms, hardware, and networks, and what really can be achieved in practice. While computing power (Moore’s law) and considerations on complexity and accuracy of algorithms were constituting the focus in this field over the last decades, new limiting factors are now encountered: The computation itself is a multifaceted process, with hardware awareness or ubiquitous parallelism due to many-core systems being just two issues emphasising this; simulation data are produced to an extent and resolution which makes their mere handling, their analysis and, thus, the extraction and representation of relevant information a serious challenge – especially if to be done interactively; the development of software for high-performance computing (HPC), ranging from specialised single-application codes to general problem solving environments, has become an increasingly complex process, comparable to what we have known for decades from other fields, and it needs sophisticated language and tool support as well as a substantial professionalisation.
Hence, to be able to exploit future (super-)computers at their full potential and to expand the research frontiers mentioned above, a concerted effort is necessary in Advanced Computing – HPC together with all its enabling technologies crucial to tackle the old and new bottlenecks. Such an endeavour must combine computing technologies and applications and bring together the HPC community with experts from the enabling technologies.