A parallel formulation for the simulation of a generic branch predictor

A parallel formulation for the simulation of a branch prediction algorithm is presented. This parallel formulation identifies independent tasks in the algorithm which can be executed concurrently. The parallel implementation is based on the multithreading model and two parallel programming platforms: pthreads and Cilk++. Improvement in execution performance by up to 7 times is observed for a generic 2-bit predictor in a 12-core multiprocessor system.