Find a way to simulate different cache architectures through premade address accessing files. This task includes modifying the associativity, the size, and the replacement policy (First In, First Out or Least Recently Used) for associativity. With this data acquired, then compare to these different influences on caches to find which provides the best solution.
As the results display, LRU always provides slightly better replacement than FIFO. However, this method also requires more power, resulting in a trade-off. These graphs also show that associativity, as expected, allows for better storage of information for caches, since it has a wider bank of choices when performing replacement. The flip-side to this is that it takes extra memory, as more tag data needs to be stored alongside the actual information.