COMP SCI 1.1

Control Unit

Coordinates activities in CPU. Directs flow of data. Completes FDE cycles, stores data in registers, sends R/W requests via control bus

Program Counter

Holds address of next instruction to be executed; next instruction in the sequence, or any branching. Increments by one.

MAR

Holds address location of data from which it is R/W from/to. Sends this via address bus.

MDR

Temporarily stores data that has been R/W from/to. Passed to it via the data bus

CIR

Holds current instruction being executed, copied from MDR where it is decoded and split into opcode and operand by the CU

ALU

Performs arithmetic and logical operations on data; bitwise and boolean.

ACC

Stores results of calculations performed by ALU.

Address bus

Carries memory addresses between MAR and RAM

Data bus

Carries data between MDR and RAM

Control Bus

Carries control signals between CU and rest of components

Clock

Unit that syncs components by generating pulses.

Interrupt Register

Checked by the CPU to see if any interrupts need processing e.g. error or keyboard inputs. If so, then Interrupt Service Routine is invoked

Cache

Small area of memory near CPU for holding frequently accessed data

FDE Cycle

PC is checked —> Address is copied into MAR —> Address sent along address bus to RAM, awaits CU signal—> CU sends signal across Control Bus —> contents sent along data bus to MDR —> Data copied into CIR —> PC incremented by 1 —> Instruction split into opcode and operand —> Instruction is executed

Pipelining

Process of carrying out multiple instructions concurrently

Von Neumann

Shared memory for data and instructions; stored in the same format. Follows linear FDE cycle. Registers are used as fast access to data.

Harvard

Instructions and data stored in separate memory units. R/W can be completed at the same time

CISC

Aims to complete tasks in as few assembly lines as possible. Multiple clock cycles per instruction

RISC

Aims to use simple instructions that will each be executed with one clock cycle

GPU

Device with thousands of independent cores working in parallel to complete repetitive tasks such as image processing. Uses SIMD to complete a single operation to multiple pieces of data in parallel

Multicore

Single CPU containing two or more independent processing units that can process multiple instructions faster than just one. However, some programs may not support multicore processing

Parallel Processing

Process of executing multiple instructions concurrently

Input device

Any device that allows the user to pass information from the outside world into a computer

Output devices

Any device that can take data in a digital format and turn it into data that a user can process

Storage device

Any device that is used for temporary and/or permanent storage of data

Optical Storage

Non-volatile secondary storage that uses a laser to R/W data by measuring the reflection of light off of pits and lands

Optical storage pros

Cheap, lightweight, portable

Optical storage cons

Slow R/W access times, prone to scratches which can corrupt data

Magnetic Storage

Non volatile secondary storage that uses magnetic fields to store binary data

Magnetic Storage pros

Cheap compared to solid state, and larger capacity

Magnetic storage cons

Slow access times, fragile and not very portable

Solid State

Non-volatile secondary storage that uses floating gate transistors to store data

Solid State Pros

Highly portable/durable. Fast access times

Solid State Cons

High relative cost. Limited R/W cycles

ROM

Small piece of read-only memory located on motherboard. Contains instructions for start-up

RAM

Temporary storage of instructions and data being executed by the processor. Volatile and faster than hard disk

Virtual Storage

Concept of storing data in a remote server that can be accessed by multiple devices from anywhere

Virtual Storage Pros

Can be accessed by anyone from anywhere

Virtual Storage Cons

Subscription fee usually has to be paid. Can only be accessed with an Internet connection