Memory management unit

A memory management unit (MMU), sometimes called paged memory management unit (PMMU),^[1] is a computer hardware unit that examines all memory references on the memory bus, translating these requests, known as virtual memory addresses, into physical addresses in main memory.

In modern systems, programs generally have addresses that access the theoretical maximum memory of the computer architecture, 32 or 64 bits. The MMU maps the addresses from each program into separate areas in physical memory, which is generally much smaller than the theoretical maximum. This is possible because programs rarely use large amounts of memory at any one time.

Most modern operating systems (OS) work in concert with the MMU to provide virtual memory (VM) support. The MMU tracks memory use in fixed-size blocks known as pages, and if a program refers to a location in a page that is not in physical memory, the MMU will cause an interrupt to the operating system. The OS will then select a lesser-used block in memory, write it to backing storage such as a hard drive if it's been modified since it was read in, read the page from backing storage into that block, and set up the MMU to map the block to the originally requested page so the program can use it. This is known as demand paging.

Modern MMUs generally perform additional memory-related tasks as well. Memory protection blocks attempts by a program to access memory it has not previously requested, which prevents a misbehaving program from using up all memory or malicious code from reading data from another program. They also often manage a processor cache, which stores recently accessed data in a very fast memory and thus reduces the need to talk to the slower main memory. In some implementations, they are also responsible for bus arbitration, controlling access to the memory bus among the many parts of the computer that desire access.

Prior to VM systems becoming widespread in the 1990s, earlier MMU designs were more varied. Common among these was paged translation, which was similar to modern demand paging in that it used fixed-size blocks, but had a fixed-size list of pages that divided up memory; this meant that the block size was a function of the number of pages and the installed memory. Another common technique, found mostly on larger machines, was segmented translation, which allowed for variable-size blocks of memory that better mapped onto program requests. This was efficient but did not map as well onto virtual memory. Some early systems, especially 8-bit systems, used very simple MMUs to perform bank switching.