Systems architecture Revision Notes

Core Components of the CPU

The Central Processing Unit (CPU) consists of three primary components. The Arithmetic Logic Unit (ALU) is responsible for all arithmetic calculations (addition, subtraction, etc.) and logical operations (AND, OR, NOT). The Control Unit (CU) manages and coordinates all parts of the CPU, directing the flow of data and instructions between memory, the ALU, and input/output devices. It decodes instructions and issues control signals. Registers are small, very fast memory locations within the CPU used to temporarily store data, instructions, and memory addresses that are actively being used by the CPU. Examples include 'Program Counter' (PC), 'Memory Address Register' (MAR), 'Memory Data Register' (MDR), and 'Accumulator' (ACC).

Von Neumann Architecture

The Von Neumann architecture is a fundamental design principle for most modern computers. It proposes that both program instructions and data are stored in the same main memory. This single address space for both instructions and data simplifies the design of the CPU and memory management. However, it can lead to a 'bottleneck' if the bus, which transfers data between the CPU and memory, becomes saturated because it is being used for both instruction fetches and data transfers simultaneously. This bottleneck is known as the Von Neumann bottleneck. Most modern general-purpose computers, including personal computers and servers, are based on this architecture.

CPU Performance Factors

Several factors influence the performance of a CPU. Clock speed, measured in gigahertz (GHz), dictates how many instruction cycles the CPU can execute per second; a higher clock speed generally means faster processing. The number of cores refers to the number of independent processing units within the CPU; more cores allow for parallel processing of tasks, improving multitasking and the performance of multi-threaded applications. Cache memory is a small, very fast memory integrated into or very close to the CPU. It stores frequently used data and instructions, reducing the time the CPU spends waiting for data from slower main memory. Larger cache sizes generally lead to better performance.

The Fetch-Decode-Execute Cycle

The CPU continuously performs the Fetch-Decode-Execute cycle to process instructions. In the fetch stage, the CPU retrieves the next instruction from main memory, determined by the Program Counter (PC), and stores it in the Instruction Register (IR). During the decode stage, the Control Unit interprets the instruction to determine what action needs to be performed, such as an arithmetic operation or a data transfer. Finally, in the execute stage, the Control Unit carries out the decoded instruction, using the ALU for calculations or interacting with other components like memory or I/O devices as required. This cycle is fundamental to how all programs run on a computer.

What is an Embedded System?

An embedded system is a computer system designed for a specific function within a larger system. Unlike general-purpose computers, embedded systems are built to perform dedicated tasks efficiently. They are not intended for programming by the end-user and run highly specialised software (firmware) that is often stored in Read-Only Memory (ROM) or flash memory. These systems are integral to many modern devices, making them 'smart' and automated.

Key Characteristics of Embedded Systems

Embedded systems are typically characterised by their single-purpose nature, meaning they perform one specific task repeatedly. They often have real-time constraints, requiring responses within a strict deadline. They are usually designed for efficiency in terms of power consumption and processing capability, often due to battery operation or thermal limits. Reliability is critical, as failure can have serious consequences. They also tend to be robust, operating in environments that might be unsuitable for general-purpose computers. Cost-effectiveness is another major aspect, driving the use of minimal resources.

Components and Examples

The core of an embedded system is usually a microcontroller or microprocessor, coupled with specific input/output components (like sensors and actuators) and memory (RAM for volatile data, ROM/Flash for firmware). Software (firmware) is burned onto the memory and is specific to the device's function. Examples include washing machines (controlling cycles), smart thermostats (regulating temperature), car engine management systems (optimising fuel and emissions), and digital watches (displaying time and basic functions). The tight integration of hardware and software is fundamental to their operation.

Clock Speed

Clock speed is a fundamental measure of CPU performance, typically expressed in gigahertz (GHz). It represents the number of processing cycles the CPU can execute per second. A higher clock speed means the CPU can process more instructions in a given amount of time, leading to faster application execution. For example, a 3.5 GHz CPU completes 3.5 billion cycles per second. While higher clock speed generally indicates better performance, it's not the only factor, as architectural differences between CPUs can mean a lower clock speed processor from a newer generation might outperform an older, higher-clock-speed one.

Number of Cores

A CPU core is an independent processing unit within the CPU package, capable of executing instructions. Early CPUs had a single core, but modern CPUs often have multiple cores (e.g., dual-core, quad-core, octa-core). Multiple cores allow the CPU to perform multiple tasks or threads of a single task concurrently. This significantly improves performance for applications designed to utilise multiple threads (multi-threaded software), such as video editing, 3D rendering, and certain games. However, for single-threaded applications, adding more cores may not yield substantial performance benefits.

Cache Memory

Cache memory is a small, very fast type of RAM located directly on or very close to the CPU. Its purpose is to store frequently accessed data and instructions, reducing the time the CPU spends waiting for data from slower main memory (RAM). There are usually multiple levels of cache (L1, L2, L3), with L1 being the fastest and smallest, and L3 being the slowest and largest. A larger cache size means the CPU can store more frequently needed data, leading to fewer accesses to main memory and consequently, faster overall processing.

Evaluating CPU Performance Factors

When evaluating CPU performance, it's essential to consider the interplay between clock speed, number of cores, and cache size. For general-purpose computing and single-threaded applications, clock speed is often a significant factor. For demanding tasks like video rendering or large data processing, having more cores becomes crucial. A larger cache benefits nearly all types of workloads by reducing memory access latency. The ideal balance depends on the specific use case of the computer. Modern CPUs are designed for efficiency, balancing these factors to offer optimal performance for a wide range of tasks.