Smart vehicles are running various computationally intensive tasks necessary for intelligent and safe driving. To save energy and enhance computational efficiency, mobile edge computing (MEC) provided by 5G networks is essential for future smart transportation. In this regard, our project investigates the problem of task offloading from multiple coexisting smart vehicles to road-side 5G vehicular edge computing (VEC) servers, while trying to capture their joint competitions in channel resource allocation and computational resource allocation. Regarding channel resource competition, we consider two types of network interfaces available in 5G, sub 6 GHz cellular bands and mmWave bands. In addition, regarding computational resource competition, we aim to let each vehicle make the best decision between local computing and offloaded computing. To derive the best offloading strategy per vehicle, we develop a game theoretic framework and determine its Nash equilibrium (and other game-related properties).