The block replacement process in the memory cache is an essential technique in computing systems to improve the efficiency of data retrieval from high-speed memory. Various caching algorithms have been developed to speed up data retrieval access in the memory cache, including Least Recently Used (LRU), Least Frequently Used (LFU), and First In First Out (FIFO). This study aims to develop a simulator by combining the LRU and LFU methods called Segmented Least Recently Used (SLRU), which is able to process data retrieval from the memory cache more efficiently. Experiments on the simulation program created were carried out on 10 random data groups to determine the effectiveness of each block replacement algorithm. Based on the test results, SLRU had the best performance, with an average hit ratio of 71.4%, followed by LRU (67%), LFU (62%), and FIFO, which showed the lowest hit ratio performance with a hit ratio of 55.8%. The advantage of SLRU lies in dividing cache segmentation into two segments: the probationary segment (LRU) and the protected segment (LFU). Based on the experiment results, it was concluded that SLRU has more efficient results in handling dynamic data access patterns than other algorithms.
Copyrights © 2025