170 likes | 373 Views
Memory. and stuff. l e well drawn pyramid. Features. Name. Description. Data is lost quickly after powering down Data can be overwritten CPU can directly access data in memory. Volatile Mutable Accessibility. Cache Memory - features. Top level of memory Is volatile Mutable
E N D
Memory and stuff.
Features Name Description Data is lost quickly after powering down Data can be overwritten CPU can directly access data in memory • Volatile • Mutable • Accessibility
Cache Memory - features • Top level of memory • Is volatile • Mutable • Accessible
Cache Memory - description • Copies data from other places • Commonly has two levels. • The first level (L1 cache) is directly ‘in’ the CPU • Smallest size (2-64kB) • Fastest speed (lowest distance to CPU) • L2 cache was formerly near the CPU on the motherboard • Relatively larger than L1 • Pretty fast too!
Cache Memory - purpose • When the CPU reaches for data, it first looks in the cache. • If it finds what it wants, it’s called a ‘cache hit’ • Else, it will reach to the other levels of memory, and this is a ‘cache miss’ • Roughly 95% cache hits [citation needed]
Cache Memory - analogy • Librarian (CPU) looks for things from book storage room • Cache is like a pouch storing most commonly borrowed books, saving the time needed to • Walk to room and back (data transfer round trip) • Search for the book in the room • Instead, Librarian first searches inside pouch containing 5 books • Smaller volume => easier to search • Virtually 0 distance
Cache Memory (tl;dr) • Top level • Volatile, mutable, directly accessible • Small, fast • Stores copies of most commonly used data
Primary Storage - features • Next level after cache. Most common example is RAM (random access memory) • Is volatile • Mutable • Accessible through memory bus
Primary Storage - description • DRAM (dynamic) • Data is stored in capacitors that act as bits • Charged = 1, no charge =0 • Capacitors leak over time • => data leaks over time too • Data has to be periodically rewritten into DRAM • ‘refresh cycle’ • Most common form (cheapest)
Primary Storage - description • SRAM (static) • Uses 4-6 transistors/bit than DRAM • DRAM uses 1 transistor + 1 capacitor/bit • Thus, costs more to make • Because of how it works, data is remanent (‘stays’) • But it is still volatile (when it loses power, data eventually vanishes) • Significantly less powerhungry • No need for data refresh • As little as a few microwattsH in idle state
Secondary storage - features • Non-volatile • Stored on tape/metallic discs • Read/write with magnetic heads (think turntable) • May be mutable or not • Not directly accessible • Uses either primary (RAM) or I/O channels
Secondary storage - purpose • Storing data permanently • Completely mutable (hard disk drive) • This type is mutable; stored on disk but can be rewritten • Partially mutable (flash memory) • Electronically rewritable • Used by USBs, cameras, PDAs, phones… • Completely unmutable (ROM) • Used to store information permanently (CDs, for example)
Secondary storage - purpose • Difference in mutability, from HDD to ROM • Speed goes up • Flash memory is about as fast as DRAM • pure ROM is as fast as SRAM • Portability goes up
Secondary storage - BIOS • The computer boots up through this • BIOS is ROM that primary memory reads from • Primary memory passes this info to CPU
Virtual Memory • Virtualization of different types of storage • Creates a virtual version of memory • The OS (and other programs) treat the computer as if there is only one type of memory – “Virtual” memory • Space is borrowed from secondary storage, speed from primary storage
Virtual Memory • “Page file” is a block of memory • Its size is the size of space borrowed from secondary storage • Thrashing occurs when you use up your virtual memory, and the computer repeatedly swaps between virtual and actual memory