Algorithm for hot paths prediction using hidden Markov model
LIU Kui1, LI Shi-ying1, LI Rui1,2, LI Ren-fa1
(1.School of Computer & Communication, Hunan University, Changsha 410082, China;2.School of Computer, National University of Defense Technology, Changsha 410073, China)
Abstract:Method of hot paths-based dynamic optimization is effective for improving the operational efficiency of the software in dynamic binary translator. This study focused on how to identify the hot paths by using the existing limited amount of previous operational information of basic blocks,and to enhance the hit rate of the prediction,with no increase of computational cost at the same time.There had been few methods based on models among exsiting hot paths prediction algorithms,which need complicated implementation.This paper proposed an improved hot paths prediction algorithm based on hidden Markov model.Since the sequence of state transition was unique,this algorithm was easy to implement,and could improve the hit rate of hot paths as well as the performance of the dynamic binary translator.The experimental results verified the efficiency of our algorithm.......