Inside the Machine

Cache Organization: Blocks and Block Frames

One way that caches take advantage of locality of reference is by loading data from memory in large chunks. When the CPU requests a particular piece of data from the memory subsystem, that piece gets fetched and loaded into the L1 along with some of its nearest neighbors. The actual piece of data that was requested is called the critical word, and the surrounding group of bytes that gets fetched along with it is called a cache line or cache block. By fetching not only the critical word but also a group of its neighbors and loading them into the cache, the CPU is prepared to take advantage of the fact that those neighboring bytes are the most likely to be the ones it will need to process next.

Cache blocks ...