This volume of the series “Advances in Parallel Computing” contains the proceedings of the International Conference on Parallel Programming – ParCo 2013 – held from 10 to 13 September 2013 in Garching, Germany. The conference was hosted by the Technische Universität München (Department of Informatics) and the Leibniz Supercomputing Centre.
With ParCo 2013, the biennial ParCo conference series now looks back at 30 years of top-level research in parallel algorithms, architectures and applications. It has finally entered an era in which parallel computing – for many years the enabling technology of high-end machines – is now ubiquitous and the key for the efficient use of any kind of computer architecture: from embedded and personal up to exascale systems.
The trend towards heterogeneous architectures, multiple levels of parallelism and towards higher and higher core numbers of supercomputing platforms, which was already addressed in the previous ParCo instances, can now be seen in full bloom. Parallel programming models for multi- and manycore CPUs, GPUs, FPGAs, and heterogeneous platforms have been one of the clear focal points at ParCo 2013. In addition, performance engineering processes, including analysis, tools and metrics, must be adapted to these new and innovative platforms. It also becomes apparent from the contributions that novel numerical algorithms are required: for basic tasks in numerical linear algebra as well as for adaptive or space-time parallel simulations. Most important, all these aspects need to be combined in the parallelisation and optimisation of large-scale applications, in order to make parallel computing – including the processing of large data sets (“Big Data”) – a persistent driver of research in many fields of science and engineering.
ParCo 2013 strongly profited from its 12 mini-symposia (including an industry session and a special PhD Symposium), which represented and intensified the discussion of current “hot topics” in high performance and parallel computing in an excellent manner. At least three mini-symposia were dedicated to large-scale supercomputing, in particular. Three mini-symposia focused on novel challenges arising from parallel architectures (multi-/manycore, heterogeneous platforms, FPGAs). A further mini-symposium hotspot was established by the “multi”-challenges: multi-level algorithms as well as multi-scale, multi-physics and multi-dimensional problems.
We would like to express our sincerest thanks to ParCo's four keynote speakers – Pete Beckman, Sudip Dosanjh, Wolfgang Nagel and Martin Schulz – who, in their presentations, gave an exciting overview of both promises and challenges for the age of exascale and Big Data. We are equally obliged to all presenters at the conference, all authors and co-authors who contributed to these proceedings, and of course to all attendees at ParCo 2013 – all of them contributed to the excellent scientific quality of the conference and to its inspiring atmosphere. Last, but definitely not least, special thanks go to all (co-)organisers, including the mini-symposium organisers, to the members of the international programme committee, and to all persons who assisted during the conference.
Gerhard R. Joubert