A1.1 Computer hardware and Operation

CPU

is the “brain” of a computer system. It is the primary computational engine of the computer, but can only process data by following instructions. It cannot actually “think” for itself. It coordinates data movement within the system

What is the CPU made of?

CU and ALU

Control Units

Directs the operations of the processor. It is responsible for the fetch - decode - execture cycle, managing all three operations and directing the computer’s memory, ALU and input/output to respond appropriately

ALU

This unit is responsible for performing arithmetic and logic operations. These include basic artithmetic operations such as addition, subtraction, multiplication and divison, as well as logic operations including AND, OR, XOR, and NOT.

Registers

very small amounts of storage that are avaliable directly on the CPU to hold temporary data that the CPU may be working on.

Instruction Register (IR)

When an instruction is fetched from memory, it is held in the IR within the CPU. This register holds the instruction that is currently being exectued by the CPU.

Program Counter (PC)

Holds the address of the next instruction that is to be fetched from memory. Once the instruction has been fetched, the PC updates to point to the next instruction that will be needed.

Memory address register (MAR)

Holds the memory address that is currently being fetched. The content from the PC is copied to the MAR, and the MAR provides this address to the memory unit, so that data and instructions can be read from or copied to that location.

Memory data register (MDR)

This holds the data that has been fetched or is about to be written to the memory address currently in the MAR

Accumulator (AC)

This stores the intermediate arithmetic or logical results produced by the ALU. 

Buses

are a critical component of a computer system, as they transfer data between various devices, including the CPU, memory, storage, and peripherals.

Control Bus

is used to transmit command and control signals from the CPU to other components of the system, and vice versa. Due to the need for signals to be sent and received, this bus is bidirectional. Some of the signals that would be transmitted via the control bus are read / write operations, interrupt requests, clock signals for synchronization and status signals from hardware components

Data bus

carries the data being processed between the CPU, memory and other peripherals. The width of the data bus is important for determining the amount of data it can transfer at one time. Common data bus widths are 8, 16, 32 and 64 bits. As data needs to be read from and written to memory, data buses are usually bidirectional. 

Address Bus

is used to transmit the address that is to be read from or written to in memory. The width of this bus determines the memory capacity of the system. For example, a 32-bit address bus can address 2³² memory locations.

Single Core Processors

This CPU is a single processing unit, meaning it can only handle one task at a time. These are more often found in low end computers or older machines. They are adequate for simple tasks that do not require heavy multitasking. Single core processors are able to run more than a single application at a time, but the CPU has to be shared between these applications, which can impact performance. 

Multi Core Processors

A CPU with multi core processors has two or more cores that can run multiple instructions simultaneously. These are often referred to as a dual core, quad core, hexa core, octa core. Their performance is signficantly faster than single core processors and they are ideal for multitasking, gaming, and servers. However, softerware has to be written to task advantage of these extra cores. Older software that does not do this would likely run at a similar speed as on a single core processor

Co Processors

A co processor is a special type of processor that has a specfic job to support the main CPU. These are built with a distinct purpose to achieve optimal performance compared to a general purpose CPU. Tasks are offloaded by the CPU to the co processor so they can in parallel, enhancing the systems performance. 

Graphics Processing Unit (GPU)

Is a specialized electronic circuit designed to accelerate the rendering of images, videos, and animations by performing rapid mathematical calculations.

Shaders and Textures

techniques used in 3D rendering to apply effects, lighting and details to models. 

Parallel processing

the ability of the GPU to perform many calculations simultaneously due to its highly parallel structure

Matrix and vector multiplications

fundamental operations in machine learning and graphics that involve complex matematical calculations. 

Deep learning

A subset of machine learning that uses an articial neural network to imitate the deign of the human brain to find generalizations in complex data that can be used to decision making

Proof of work

a consensus mechanism requiring cryptominers to solve complex problems to add a new block to the blockchain

Rendering

the process of generating an image from a model by means of computer programs

Vertex and pixel data

data used by the GPU to render 3D objects and images

Frame

a single image in a sequence of images that makes up a video or animation

Volatile memory

Requires power to maintain the stored data. When the power is turned off, the data is lost. An example of volatile memory is RAM (random-access memory).  

Non-volatile memory

Retains stored information even when the power is turned off. Examples include ROM (read-only memory), SSDs (solid state drives), and HDDs (hard disk drives). 

Random Access Memory (RAM)

Holds instructions and data for programs that are currently running.

Read Only Memory (ROM)

Used for storing instructions that are very rarely modified.

Cache hit

When the CPU requests data and it is found in the cache memory

Cache miss

When the CPU requests data and it is not found in the cache memory, necessitating retrieval from sloewr main memory or storage

L1 Cache

is the smallest and fastest, like the very centre of the onion, where everything is tightly packed and closest to the core of the CPU

L2 Cache

is slightly larger and slower, like the next layer out — still close to the centre, but not as quick to access as the very core

L3 Cache

is the largest and slowest, like the outer layers of the onion. It’s still important, but it takes a bit longer to get to, just like how the CPU takes a bit more time to access data in L3 cache compared to L1 and L2.

Fetch

The CPU fetches an instrcution from the memory

Decode

The CPU interprets the instruction and prepares the necesssary operations to execute it

Execute

The CPU performs the actions required by the instruction

Multi-core archiectures

systems with multiple CPU cores on a single chip, allowing parallel execution of instructions and tasks

Fetch-Decode-Execute Cycle

The process the CPU follows to execute instructions from memory.

Latency

is defined as the Time required for an instruction to propagate through the pipeline. Its unit is a time unit like microseconds, nanoseconds.

Throughput

 is the rate at which instructions can start and finish i.e. the number of instructions finished per second.

Normally Cycles per Instruction (CPI)

is used as an abstract measure in many calculations for comparison between  CPUs

Hard Disk Drive (HDD)

are older technology but are often still used, especially in non mobile devices, as they are relatively cheap compared to the amount of storage they offer. They utilize a spinning magnetic disk to read/write data. They are suitable for storing large volumes of data, such as media files, backups and documents, where speed is not critical

Solid State Drive (SSD)

Have no moving parts. They use flash memory to store data, offeringhigh speed data acess and durability. This makes themvery popular in portable devices such as laptops and tablets. They are ideal for operating systems, software applications and games due to their fast read / write speed, which enhances the overall system performance.

Embedded MutliMediaCard (eMMC)

A type of flash storage that utlizies NAND flash memory. They are soldered directly on to the motherboard of the device.

Network Attached Stoage (NAS)

A dedicated file storage connected to a network that allows multiple users to access data.

Redundant Array of Independent Disks (RAID)

a data storage technology that combines multiple physical drives into a single logical unit to improve performance, provide redunddancy and ensure data protection

Loseless compression

When data is compressed to a smaller size, but can be restored back to the original without any loss of information.

Statistical redudancy

the repetition of information within a data set that does not contribute to its uniqueness

Run Length Encoding (RLE)

Run length encoding is an effecive loseless data compression technique used to reduce the size of files containing many consecutive repeated characters.

Low frequency data

Correspond to slow changes in pixel values, such as broad areas.