Cache management policy in gem5
The rapid development of computing platforms has widened the gap between the computing system and memory system, placing more pressure on cache, which is an integral part of the memory system. Despite numerous studies on cache management policies to optimize resource usage, some of them cannot keep up with the fast-paced trends in computing devices. Many of the state-of-the-art cache replacement policies and prefetchers in our research group are based on simulators with simple hardware abstraction for easy development and prototyping, but they do not support more realistic environments, such as cache coherence and heterogeneous systems. This thesis aims to experimentally transplant several cache management policies to more advanced simulators and provide initial experience in dealing with the challenges encountered in the process.