作者: Edwin J. Lau
DOI: 10.1016/0166-5316(82)90011-6
关键词:
摘要: Abstract Prefetching is a technique applied to memory management policies in which pages are brought into before they actually needed. In this study, prior knowledge of program behavior obtained through trace data used parameterize variation the working set policy supporting demand page prefetching using one lookahead. Two new algorithms double proposed. A comparative analysis them made. Evaluation techniques driven simulation shows that, general, very effective reducing fault rate and some cases space time product. As expected, occupancy increases, but usually by small amounts. The results encouraging for their ability improve performance variety programs exhibit form locality (spatial and/or temporal).