Overview
No single memory technology is simultaneously fast, large, and cheap. Computers instead use a
RAM Fundamentals
"RAM" isn't one technology — it's a family. The RAM that makes up multi-gigabyte main memory
CPU Caches
A CPU cache is a small amount of fast SRAM that sits between the core and main memory, holding
Virtual Memory & Paging
Every process behaves as if it owns the entire address space and as if its memory is one large,
The TLB
The cache that makes virtual memory affordable: TLB structure, page-walk caches, address-space tags, and shootdown cost.
Cache coherence
How several caches holding the same line stay consistent, and why two unrelated variables in one line can destroy performance.
NUMA
When "main memory" stops being one thing: node distance, local versus remote latency, and what interleaving trades away.