I finished my PhD in Astrophysics in October 2013 at Sterrewacht Leiden, under supervision of Prof. Simon Portegies Zwart and Prof. Cees de Laat (University of Amsterdam). Subjects addressed in my PhD thesis include high-resolution dark matter simulations (CosmoGrid) and simulations of star clusters within a larger-scale environment.
My PhD thesis is available for download at the library of Leiden University.
The CosmoGrid simulation is a LambdaCDM simulation containing 2048^3 particles in a cube with sides of 30 Mpc (at z=0). It consists of 552 snapshots, taken between z=65 and z=0. The CosmoGrid simulation was completed on multiple supercomputers running in parallel, each computing part of the volume while communicating to the others. We used the Huygens supercomputer in Amsterdam, the Cray XT4 in Tokyo, HECToR in Edinburgh and Louhi in Espoo. We showed that a large simulation can run on multiple supercomputers in parallel without being hindered by the communication between these computers. Compared to a run on one site, we obtained an efficiency of 93% when running on three supercomputers. The work related to the technical setup of CosmoGrid is published in Portegies Zwart et al. (2010), Groen et al. (2010) and Groen et al. (2012). Astrophysical results resulting from the CosmoGrid simulation have been published in Ishiyama et al. (2011), Rieder et al. (2013a) and Rieder et al. (2013b).
For most of the simulations I performed during my PhD project, I used the AMUSE environment (Portegies Zwart et al. 2012, Pelupessy et al. 2013). Amuse is a Python layer that allows the user to run a simulation using a simple Python script, the communication to and between the simulation codes (such as a direct N-body code or tree code and a stellar evolution code) is then handled by AMUSE.