[matthias509]

I’m guessing you are thinking of cache lines where CPUs will read/write data in 64 byte chunks from memory to/from cache.

As you said, being aware of this lets you optimize away cache misses by controlling the memory access pattern.