Theoretical modeling of plasmonic phenomena is of fundamental importance for rationalizing experimental measurements. Despite the great success of classical continuum modeling, recent technological advances allowing for the fabrication of structures defined at the atomic level require to be modeled through atomistic approaches. From a computational point of view, the latter approaches are generally associated with high computational costs, which have substantially hampered their extensive use. In this work, we report on a computationally fast formulation of a classical, fully atomistic approach, able to accurately describe both metal nanoparticles and graphene-like nanostructures composed of roughly 1 million atoms and characterized by structural defects.
