Modeling across physical domains can lead to coupled systems that operate on different time scales. Moreover, spatially discretized models can have dynamic behavior on only a small portion of the whole system, while that portion might move over time. With a single rate solver like DASSL, the step size is restricted to fast dynamics of a few states while the others could do much larger time steps, degrading the overall simulation speed. Recent development of the Generic Bi-rate ODE (GBODE) solver within the OpenModelica simulation environment addresses these models. It allows for adaptive partitioning of slow and fast states and progresses these partitions with different time steps. However, substantial performance improvements were not yet observed, in part because of the rigid evaluation structure of the generated code. During each small step only the derivatives of fast states are required, but all derivatives are computed. This paper demonstrates the use of selective equation evaluation to reduce the computational cost of multi-rate integration as shown on a scalable example of a distributed heating system. Further uses for selective evaluation during the simulation process are discussed and a complexity analysis is given.