The invention is a technique that allows very efficient execution of an instruction set simulator (ISS) when certain code fragments are simulated. They occur frequently in operating systems, when a processor waits for something to occur. These loops perform very little, if any, computation. An automatic analysis of the code fragments is performed to determine whether they can be transformed to “fast forward” when execution loops through the code fragment. This is done by the creation of a function g, that takes a number of state changes as an input and can compute the state at a time T by a single application of the function, rather than applying the next state function repeatedly until time T is reached. An important aspect of the resulting optimization is that it produces exactly the same simulated result as compared to when it is not applied, i.e. the optimization is non-intrusive.