There are many astrophysical and laboratory scenarios where kinetic effects play an important role. These range from astrophysical shocks and plasma shell collisions, to high intensity laser-plasma interactions, with applications to fast ignition and particle acceleration. Further understanding of these scenarios requires detailed numerical modelling, but fully relativistic kinetic codes are computationally intensive, and the goal of one-to-one direct modelling of such scenarios and direct comparison with experimental results is still difficult to achieve. In this paper we discuss the issues involved in performing kinetic plasma simulations of experiments and astrophysical scenarios, focusing on what needs to be achieved for one-to-one direct modeling, and the computational requirements involved. We focus on code efficiency and new algorithms, specifically on parallel scalability issues, namely on dynamic load balancing, and on high-order interpolation and boosted frame simulations to optimize simulation performance. We also discuss the new visualization and data mining tools required for these numerical experiments and recent simulation work illustrating these techniques is also presented.