摘要
针对LIRS算法结构复杂、适应性差的不足,提出一种改进的自适应时钟算法。在LIRS算法基础上借鉴Clock算法思想,简化算法组织结构,加强对不同数据访问模式的适应性和捕获高频数据的能力。实验结果证明,与LIRS算法及2种传统替换算法相比,自适应时钟算法在命中率上有更好的表现,较好地满足用户对Cache高命中率和占用额外计算资源少的要求。
To solve the problem that LIRS algorithm has the shortage of complex structure and poor adaptability, this paper proposes an Adaptive Clock algorithm based on LIRS algorithm. Adaptive Clock algorithm absorbs the Clock algorithm ideas to simplify the organizational structure of the algorithm to improve the adaptability of different data access patterns, as to enhance the ability to capture high-frequency data. Experimental results show that Adaptive Clock has better performance in hit rate than LIRS algorithm and two commonly used algorithms, and meets the requirements of the high Cache replacement ratio and less computing resource.
出处
《计算机工程》
CAS
CSCD
2012年第20期286-289,共4页
Computer Engineering
基金
国家自然科学基金资助项目(61003036
60873138)
黑龙江省教育厅科学技术研究基金资助项目(12513048)