The B&Mode Branch Predictor

Dynamic branch predictors are popular because they can deliver accurate branch prediction without changes to the instruction set architecture or pre-existing binaries. However, to achieve the desiredprediction accuracy, exist-ing dynamic branch predictors require considerable amounts of hardware to minimize the integerence effects due to aliasing in the prediction tables. We propose a new dynamic predictor, the bi-mode predictor, which divides the prediction tables into two halves and, by dynamically deter-mining the current “mode ” of the program, selects the ap-propriate half of the table for prediction. This approach is shown to preserve the merits of global history basedpredic-tion while reducing destructive aliasing and, as a result, im-proving prediction accuracy. Moreover, it is simple enough that it does not impact a processor’s cycle time. We con-clude by conducting a comprehensive study into the mech-anism underlying two-level dynamic predictors and investigate the criteria for their optimal designs. The anal-ysis presented provides a general framework for studying branch predictors. 1