Load Value Prediction Method to Utilize Prediction Miss Bias

Abstract

Data hazard caused by load instruction is a serious problem for superscalar processors to improve instruction level parallelism. Various load value prediction methods have been proposed to reduce the data hazard. If the prediction is correct, successive instructions can be executed in parallel without waiting for memory access. However, prediction miss… (More)

Topics

6 Figures and Tables

Cite this paper

@inproceedings{Moriwaki2015LoadVP, title={Load Value Prediction Method to Utilize Prediction Miss Bias}, author={Nobuhiro Moriwaki and Lin Meng and Shigeru Oyanagi}, year={2015} }