缓存是增强计算机整体性能的一项重要技术，缓存替换算法作为核心技术尤为重要。通过深入研究多种数据访问模式下 ARC（adaptive replacement cache）算法的性能，总结出 ARC 算法性能较差的数据访问模式及其原因。针对发现的不足提出了一种改进的缓存替换算法，该算法通过引入 IRR（inter reference recency）信息，提高了弱局部性访问模式下的缓存命中率，改善了 ARC 算法对不同数据访问模式的适应性。仿真实验结果表明，改进后的算法提升了应对弱局部性访问模式的能力，增强了算法的灵活性。%Cache is an important technology to improve the overall performance of computers.As its core technology,more impor-tance is attached to cache replacement algorithm.Various data access patterns have been thoroughly studied,and the patterns that ARC (adaptive replacement cache)algorithm handles poorly and its cause are summed up.According to the weakness,an improved cache replacement algorithm is proposed.By introducing IRR (inter reference recency)information,the new algorithm improves cache hit rate of weak locality access pattern and makes ARC algorithm more adaptive.The simulation result shows that this algorithm can improve the ability to handle weak locality access pattern and enhance the flexibility of ARC algorithm.
展开▼
