Aim of the project is the execution of computationally extremely challenging CFD simulations of blood flow and contrast agent (CA) transport in coronary arteries on HPC clusters. Therefore scalability testing is performed to assess the suitability of applied software codes in HPC environments. Cardiovascular 3D-models are extracted from high-resolution cryomicrotome imaging data and meshed with computational grids. Navier-Stokes-equations for blood flow and the advection-diffusion-equation for CA transport are solved to obtain CA bolus dispersion in the epicardial vessels. We find indications of asymptotically decreasing dispersion effects with increasing vessel generation.