The dynamic properties of a granite rock employed in crushed form as coarse aggregate in high-strength concrete, were experimentally characterised, and the average values obtained were assumed to be representative of the properties of the coarse aggregate. Cylindrical specimens were core-drilled from the granite rock, then cut into various lengths for different tests, i.e. static and dynamic uniaxial compression, as well as static and dynamic splitting tension (Brazilian tests). From the data obtained, the effects of strain rate on the mechanical properties (e.g. stress-strain relationship, tensile and compressive strengths) were analysed. Significant rate dependence was observed for both tensile and compressive responses, and this was quantified via a dynamic increase factor (DIF, the ratio between the dynamic and static strength). For compression, the DIF follows the CEB-FIP 2010 equation frequently used to describe the rate sensitivity of normal-strength concrete. The tensile response shows a stronger rate dependence than the compressive response, and its DIF deviates from predictions based on the CEB-FIP equation, which underestimates the tensile DIF, and defines a much higher strain rate at which transition from a gradual to a rapid increase in the DIF occurs. An empirical DIF formula, derived from fitting of experimental data, is proposed to describe the tensile rate sensitivity. Apart from experimental characterisation, efforts were also directed at identifying a constitutive model to describe the dynamic behaviour of the granite coarse aggregate, and the Johnson-Holmquist 2 (JH-2) model, commonly used to capture the impact response of brittle materials (e.g. rock, ceramic) was chosen. The dynamic tensile and compressive properties obtained from experiments, were employed to determine the material parameters in the JH-2 model. Finite element modelling of the dynamic compression and splitting tension tests was performed, using the JH-2 model and the material constants derived, to describe the behaviour of the granite specimen. The simulation results correlate well with experimental measurements.