This work investigates the potential of direction-correlations to improve branch prediction. There are two types of direction-correlation: affectors and affectees. This work considers for the first time their implications ...