Description and Aim Computational astrophysics is a fast growing discipline which envelopes computational science, observational astronomy and theoretical physics. Many recent exciting advances in astrophysics are realized through computation. With computers it has become possible to perform experiments that cannot practically be executed in a laboratory, or it is possible to advance time or to descend from an intergalactic scales down to planetary scales. The conceptual, theoretical and methodological foundations necessary to understand these multi-scale processes are dramatically limited by our ability to translate our physical understanding to computer instructions. Recent advances in our understanding of self gravitating systems, like the universe, galaxies, star clusters and planetary systems, has strongly been driven by the continuous increase in computer capacity and our ability to utilize the available hardware via software. These advances have progressed more-or-less independently in the four main astrophysical domains by the development of relatively separate numerical techniques, which include: symplectic methods, direct N-body, tree-codes and FFT-mesh-based methods. At the same time computer architectures have changed, and each of the astrophysical specialized solvers run efficiently on different hardware, like GRAPE, Graphical Processing Units, shared memory supercomputer and distributed memory Linux clusters. Regardless of the recent improvements in our understanding of the Universe it remains painstakingly clear that a proper appreciation of the complexity at all scales requires the proper modeling on all scales. But this is dramatically hampered by computer hardware and by the complexity of the software. The next generation of computer assisted astrophysics is driven by the development of advanced multiscale and multi physics software via the hybridization of the various techniques. For this reason we bring together a wide range of gravitational astrophycisists and computer scientists.Q Advanced school Objective of the advanced school We aim at educating the next generation key researchers in computational gravitational dynamics. Most individual universities lack the specialized expertise required to model gravitational dynamics on all scales relevant for understanding the Universe. During this school the apprentice will work in close collaboration with peers under the supervision of the key researcher in the field. This small splinter group will conduct a research project which should culminate in a publication. The school is meant to last for two weeks, with supervised research during the first week and independent research in the second. In the second week the apprentice has the opportunity to attend the workshop on computational gravitational dynamics at the Lorentz Center. School format Students will work in triplets under the supervision of the key researcher. Together they will work for one week with the possibility to finish the work in the second week. The results of the project will be presented at the workshop in the second week and published in a peer reviewed proceedings. Workshop Objective of the workshop We aim at bringing together a wide range of computational gravitational dynamicists to discuss the recent progress and possibilities for interdisciplinary research in this field. We notice that planetary dynamicists rarely interact with cosmologists on the numerical details used in their scientific production software, even though the underlying physics is quite comparable. We intend to bridge this gap. Format of the workshop We will have review talks of a number of invited key researchers, and short contributed presentations of cross disciplinary researchers, followed by extensive discussions.