Hardware

The Central Processing Unit (CPU) is the brains of the computer. The purpose of the CPU is to process data and instructions. Without the CPU, a computer would not be able to compute.

If a device has a button it probably has a CPU (and is therefore a computer). They carry out tasks very quickly: the CPU in your smartphone can carry out over 1 billion calculations a second.

History

Computers like Colossus, while electrical, and stored data electrically, were programmed mechanically using wires and switches. If you wanted to change the program being run on the computer, it would have to be re-wired. Imagine having to rewire a computer to swap from Word to Chrome (or from Fortnite to Minecraft)!

Computers like Colossus, while electrical, and stored data electrically, were programmed mechanically using wires and switches and if you wanted to change the program being run on the computer, it would have to be re-wired.

This was a major problem, as everytime we needed the computer to do something different it required a rebuild.

Von Neumann, a Hungarian engineer and scientist, proposed storing the program (instructions – what tells the computer what to do) electrically along with data. 

His implementation of a “Stored Program Computer” is called Von Neumann Architecture, and is the basis of the architectures computers use to this day.

Von Neumann suggested the following architectrue, where instructions are stored in memory. On modern day computers this memory could be RAM or Secondary storage, however in this architecture refers to a memory store closer to the concept of RAM.

Parts described in the architecture

During the Fetch Decode Execute Cycle data and addresses need to move around parts of the CPU and between the CPU and memory, this is achieved using buses.

Buses are tiny wires built into the CPU and other components that allow data, instructions, and signals to move around.

There are 3 main buses: 

Directionality

Both the data bus and control bus are bidirectional, which means data can be sent and recieved over the bus in both directions. The address bus is unidirectional, meaning that addresses can be sent from the CPU to memory, but not back in the reverse direction.

The Fetch Decode Execute Cycle is the only thing that a CPU can do. This process begins as soon as a computer is switched on and will only end when it is turned off.

Each step in the Fetch Decode Execute cycle makes use of different parts of a CPU and different busses.

Fetch

All the data and the instructions the CPU needs is stored in Memory (which is outside of the CPU). All the data and instructions have their own unique address for the “memory location” they are in.

When the CPU needs data or instructions it requests them using the address of their memory location. The CPU places the memory location address on the Address Bus.

The data or instruction at the memory location address requested by the CPU is sent back to the CPU using the Data Bus. The data or instruction requested is now stored in registers within the CPU.

Decode

The Control Unit inside the CPU figures out what to do with the instructions that has been fetched from memory.

Every CPU has an Instruction Set: A list of every instruction it can carry out. Each instruction has a unique binary number

Execute

The execute stage is where the instruction is carried out. This could be many things: It could be for the ALU to carry out a calculation, to fetch data, fetch another instruction at another memory location, or even to store data in memory.

Once the execute stage has been completed the cycle starts again. 

Primary Memory is a computer's short term memory. Any applications running on a computer will be transferred to Primary Memory. Once an application is no longer running it will be removed from memory.

We refer to the Primary Memory is computers to as RAM (Random Access Memory). RAM is Volatile this means that in order to store data it needs power.

Upgradability

On some computers RAM can be upgraded. On newer computers and smaller devices RAM is often soldered directly to the motherboard and cannot be upgraded.

Errors

Sometimes if a computer has an issue turning it on and off solves the issue because it clears the RAM where the issue may have been.

Why do we need RAM?

The CPU processes instructions and executes instructions. We must get the data and instructions to the CPU, and if we can’t do this fast enough the CPU will “idle” and not have anything to do.

Instructions and data are stored long term on secondary storage but accessing them directly from secondary storage can be a slow process. RAM on the other hand is much faster than secondary storage, but cannot store as much data.

Pre-loading the data and instructions into RAM allows the CPU to access them faster and can reduce possible bottlenecks. 

Secondary Storage devices are used for long term storage of data.

All secondary storage devices are non-volatile, which means that they don't require power to continue to store data. They typically store significantly more data than primary memory, but are much slower for reading / writing data.

All secondary storage devices can be categorized into one of three classes, each has their own advantages and disadvantages:

Optical

Magnetic

Solid State / Flash

General Purpose Computers are those that can have many purposes and their purpose can be changed

Desktops, Laptops, and Smartphones are all general purpose computers - we can change the software they run easily. They can be used for playing games, browsing web pages, completing office work, watching videos

These only make up 2% of computers worldwide!

Embedded Systems

Embedded Systems are computers embedded within other devices

They have a narrow or singular purpose.

Examples include:

• Washing Machines
• Microwave
• Thermostats
• Bluetooth Speaker / Headphones
• Smart Home Devices

Embedded systems are highly specialised:

• Single or limited purpose – they usually only do one thing. Due to having a single purpose many do not have operating systems, instead we call the software that runs on them “Firmware”. Some Embedded systems are have to operate in real time.
• Less processing power - you don’t need GHz to run temperature checks on a thermostat
• Smaller - less memory, smaller CPUs, less complex parts means they can be physically smaller
• Reliable - fewer parts, less complexity means they’re less likely to fail and crash
• Use less energy - lower complexity, slower speeds means needing less power to run and they run cooler (they can every run on small batteries)
• Cheaper - all the above leads to much lower manufacturing costs