Comparison of Cache Page Replacement Techniques to Enhance Cache Memory Performance

Abstract

Caching is a very important technique for improving the computer system performance, it employed to hide the latency gap between memory and the CPU by exploiting locality in memory accesses. In modern computer architectures a page miss cause the several hundred CPU cycles [1, 15]. In a memory hierarchy, a cache performs faster than auxiliary storage, but is more expensive. Some important page replacement policies such as FIFO, LRU, RANDOM and SECOND CHANCE are used in cache memory to replace the page whenever it is needed. So page replacement policy severely affects the performance of cache memory. So here our purpose is to simulate FIFO, LRU, RANDOM and SECOND CHANCE policies and to compare the results for various applications such as bzip, swim and gcc traces (taken from SPEC2000 benchmark for simulation) etc. using MS-Excel [3, 10, 12] .