Chapter 8: Caching | 🗒️ Ben's Notes

At this point, we’ve solved nearly all of the problems with base and bound memory translation. But there is one major problem left— all of this additional complexity adds lots of memory accesses, which might make things very inefficient! This is where caching comes in. Since programs only care about a very small subset of the total information available, if we identify this subset and place it into more local memory, then we can efficiently perform most memory accesses. Caches give the benefits of speed from the fastest types of memory while still having the capacity of the largest memory. Ideally, caches are implemented such that they are invisible to programmers: they will automatically move memory between locations. Caches have tags (addresses) and blocks (aligned words that store information). In order to simplify the comparison process (so that we don’t need to check every tag to see if we have a hit or miss), we can split the cache into sets. So we only need to compare a tag to a

At this point, we’ve solved nearly all of the problems with base and bound memory translation. But there is one major problem left— all of this additional complexity adds lots of memory accesses, which might make things very inefficient! This is where caching comes in. Since programs only care about a very small subset of the total information available, if we identify this subset and place it into more local memory, then we can efficiently perform most memory accesses. Caches give the benefits of speed from the fastest types of memory while still having the capacity of the largest memory. Ide