Cache prefetching in memory management greatly relies upon effectiveness of prediction mechanism to fully exploit available resources and for avoiding page faults. Plenty of techniques are available to devise strong prediction mechanism for prefetching but they either are situation specific (Locality of reference principle) or inadaptable (Markovian model) and costly. We have proposed a generic and adaptable technique benefiting from past experience by employing hybrid of Case Based Reasoning (CBR) and Neural Networks (NNs). Here we will be concerned with improving adaptation phase of CBR using NN and its impact on predictive accuracy for prefetching. The level of predictive accuracy attained (specifically in case adaptation of CBR) is ameliorated by handsome margin with declined cost than contemporary techniques as would be affirmed by results. © 2010 Springer-Verlag.
CITATION STYLE
Sarwar, S., Ul-Qayyum, Z., & Malik, O. A. (2010). CBR and neural networks based technique for predictive prefetching. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6438 LNAI, pp. 221–232). https://doi.org/10.1007/978-3-642-16773-7_19
Mendeley helps you to discover research relevant for your work.