Hardware

The CPU, microprocessors and embedded systems

The central processing unit (CPU) is the part of a computer that processes the instructions and the data that are input into the computer so that the result can be output. It carries out every instruction in a program, one after another, and it is often called the brain of the computer. A microprocessor is a type of integrated circuit on a single chip, which is what makes it possible to build a complete CPU into a phone, a washing machine or a car.

A general purpose computer, such as a personal computer or a laptop, can run many different programs and be used for many different tasks. An embedded system is a computer built into a larger device to perform a dedicated function: one job, carried out over and over. Embedded systems are found in domestic appliances such as washing machines and microwaves, in cars (engine management and braking systems), in security systems, in lighting systems and in vending machines.

Embedded systems usually have limited processing power and memory, because they only have one job to do, and they often have no keyboard or screen, or only a very simple one. The program that runs the system is normally held in ROM so that it is still there when the power is switched off, and the system is expected to keep working reliably for years without being updated.

Von Neumann architecture: units, registers and buses

In a computer with a Von Neumann architecture, the instructions and the data are both held in the same memory (RAM) and are fetched and processed one at a time by the CPU. The CPU itself is made up of units, registers and buses.

The arithmetic logic unit (ALU) carries out the arithmetic operations, such as adding and subtracting, and the logical operations, such as comparing two values. The control unit (CU) controls the flow of data and the execution of instructions: it decodes each instruction and sends the control signals that tell the other components what to do.

The registers are tiny, very fast storage locations inside the CPU. The program counter (PC) holds the address of the next instruction to be fetched. The memory address register (MAR) holds the address of the memory location that is currently being read from or written to. The memory data register (MDR) holds the data or instruction that has just been fetched from memory, or that is about to be written to memory. The current instruction register (CIR) holds the instruction that is currently being decoded and executed. The accumulator (ACC) holds the result of the calculations carried out by the ALU.

The buses are the sets of wires that carry data, addresses and signals between the components. The address bus carries the address of a memory location from the CPU to memory, and it works in one direction only. The data bus carries the data and the instructions between the CPU and memory, in both directions. The control bus carries the control signals from the control unit to the other components, such as the signal to read from or to write to memory.

The fetch-decode-execute cycle

The CPU works through a program by repeating the fetch-decode-execute (FDE) cycle. Every instruction goes through all three stages before the next instruction starts.

Fetch. The PC holds the address of the next instruction, and that address is copied into the MAR. The instruction stored at that address in RAM is fetched and placed in the MDR. The instruction is then copied from the MDR into the CIR, and the PC is incremented so that it now holds the address of the following instruction, ready for the next cycle.

Decode. The control unit decodes the instruction held in the CIR. The instruction is split into the operation that has to be carried out and the data or address it applies to, so the CU works out what has to happen and which components will be needed.

Execute. The instruction is carried out. The ALU may perform a calculation and store the result in the accumulator, data may be read from or written to memory, or the program may jump to a different instruction by loading a new address into the PC. When the instruction has been executed, the cycle begins again.

• Fetch: the address in the PC is copied to the MAR, the instruction is fetched from RAM into the MDR, it is copied into the CIR, and the PC is incremented.
• Decode: the control unit works out what the instruction in the CIR means and which components are needed to carry it out.
• Execute: the instruction is carried out, for example a calculation in the ALU with the result stored in the accumulator.
• The cycle then repeats for the next instruction, using the address now held in the PC.

Cores, cache, clock speed and the instruction set

Three things about the CPU itself affect how quickly it performs. The clock inside the CPU produces regular pulses that drive the FDE cycle, and the clock speed, measured in gigahertz (GHz), is the number of cycles the CPU can carry out each second. A higher clock speed means more instructions are processed every second, so the performance is better.

Cache is a small amount of very fast memory built into the CPU that stores the instructions and data that are used most often. When the CPU needs something, it looks in the cache first. A larger cache means more of what is needed is already close to hand, so there are fewer fetches from the much slower RAM, and the performance improves.

A core is a complete processing unit inside the CPU, so a multi-core processor can carry out several instructions at the same time. More cores mean more instructions can be processed simultaneously, which improves performance, provided the software has been written to make use of more than one core.

An instruction set is the list of all the commands that can be processed by a CPU. Each command in the set is written in machine code for that particular processor. Different processors have different instruction sets, which is why a program compiled for one type of CPU will not run on another type without being recompiled.

Input devices, output devices and sensors

An input device is hardware that is used to enter data or instructions into a computer system. Without an input device the computer would have nothing to process, because the program and the data both have to get into the system somehow. An output device is hardware that presents the result of the processing, whether as an image on a screen, as printed text, as sound or as a physical movement.

The input devices you need to know include the barcode scanner, which reads the bars and spaces using reflected light and converts them into a code that is looked up in a database, and the QR code scanner, which reads a two-dimensional code that can hold far more data than a barcode, such as a web address. A digital camera captures an image on a sensor and converts it into a digital file, a keyboard enters text and numbers, and a microphone converts sound waves into a signal that the computer samples and stores as binary. An optical mouse uses a light and a sensor to detect movement by comparing successive images of the surface underneath it. A 2D scanner scans a flat document and produces a digital image, while a 3D scanner scans an object and produces a digital 3D model that can then be printed.

Touch screens work in three different ways. A resistive touch screen has two conductive layers that touch when the screen is pressed, so it responds to a finger, a stylus or a gloved hand, but it is less sensitive and usually detects only one touch at a time. A capacitive touch screen detects the change in capacitance caused by a finger touching it; it is more sensitive and supports multi-touch, but it needs a bare finger and will not respond to a glove or a plastic stylus. An infra-red touch screen has a grid of infra-red beams across it, and a touch breaks the beams so that the position can be worked out; it works with any object.

The output devices you need to know include the inkjet printer, which sprays tiny droplets of ink onto the paper and is cheap to buy and good for colour photographs, but slow with expensive ink; the laser printer, which uses a laser to build up an image on a drum before toner is transferred to the paper and fused with heat, and is fast with a low cost per page in large volumes; and the 3D printer, which builds up a solid object layer by layer from a digital model and is used for prototyping. An LED screen uses light emitting diodes for each pixel, so it is bright and energy efficient, while an LCD screen uses pixels of liquid crystal that block or let through the light from a backlight, so it is thin, light and low power. A DLP projector uses a chip of thousands of tiny mirrors to switch light on and off, giving a smooth image with no visible pixel grid, and an LCD projector shines light through liquid crystal panels and a lens. A speaker converts an electrical signal into sound waves, and an actuator converts a signal from the computer into a physical movement, such as a motor turning a valve or a robotic arm moving.

A sensor is a device that automatically captures a physical quantity from its surroundings and converts it into a digital signal that a computer can process. Sensors are used so that a system can monitor its environment and react without a person having to take readings and type them in. Each sensor measures one thing, and the computer compares the value it receives with a set range and takes action, for example switching on a fan, opening a valve or sounding an alarm.

Primary storage, secondary storage and virtual memory

Primary storage is storage that is directly accessed by the CPU. It holds the instructions and the data that are being used at that moment. RAM (random access memory) is volatile: it can be read from and written to, it holds the programs and data currently in use, and its contents are lost when the power is turned off. A computer with more RAM can keep more programs and more data open at once without slowing down.

ROM (read only memory) is non-volatile: its contents are not lost when the power is turned off, and it can be read but not changed. It holds the start-up instructions, often called firmware, that the computer needs in order to boot up and check its hardware before the operating system can be loaded from secondary storage. A computer needs both: ROM holds the instructions that must be there every time the computer is switched on and must never be lost, while RAM is the fast, changeable workspace the CPU uses while the computer is running.

Secondary storage is not directly accessed by the CPU, and it is needed for more permanent storage of data. It keeps files, programs and documents when the computer is switched off, and it holds far more data than RAM for a much lower cost, although it is slower to access. There are three types you must know.

Magnetic storage uses platters that are divided into tracks and sectors, and data is read and written using electromagnets that magnetise small areas of the surface. A hard disk drive (HDD) is the main example. Magnetic storage gives a very large capacity at a low cost per gigabyte, but it has moving parts, so it is slower than solid-state storage and can be damaged if the drive is knocked while it is working.

Optical storage uses lasers to create and read pits and lands on the surface of the disc: the laser is reflected differently from a pit and from a land, and that difference is read as 1s and 0s. CD, DVD and Blu-ray discs are the examples. Optical media are cheap and portable and are good for distributing files, but they hold less than a hard disk drive, they are slower, and the surface can be scratched.

Solid-state (flash memory) storage uses NAND or NOR technology, in which transistors act as control gates and floating gates to trap charge and so store bits. A solid-state drive (SSD), an SD card and a USB drive all work in this way. Because there are no moving parts, solid-state storage is fast, silent, uses little power and is not damaged by being moved, which is why it is used in phones and tablets. It costs more per gigabyte than magnetic storage, and each cell can only be written to a limited number of times.

Virtual memory is a section of secondary storage, usually on the hard disk drive or the SSD, that the computer uses as an extension of RAM. When RAM is full, pages of data are transferred between RAM and virtual memory as they are needed, so the computer can keep more programs open than the RAM alone would allow. It is necessary because RAM is limited and the programs and data in use need more space than is available. The drawback is that reading from and writing to secondary storage is much slower than reading from RAM, so a computer that relies heavily on virtual memory slows down noticeably.

Cloud storage and network hardware

Cloud storage means storing data on remote servers that are owned and managed by a cloud provider, rather than storing it locally on your own device. The data is uploaded over the internet and accessed over the internet, so physical servers and storage are still needed, they are simply not yours.

The advantages are that the data can be accessed from any device with an internet connection, so files can be shared between devices and with other people; the provider manages the hardware, the backups and the security updates, so there is no need to buy, replace or maintain storage hardware; the storage can be increased easily when it is needed; and the files are safe if your own device is lost, stolen or damaged. The disadvantages are that an internet connection is needed to reach the files at all, so they cannot be used offline; accessing a large file can be slower than reading it from a local drive; there is usually an ongoing subscription cost; and the data is held by a third party, so there are security and privacy risks and you lose direct control of it.

A computer needs a network interface card (NIC) to access a network. The NIC is the hardware that connects the device to the network and converts the data into a form that can be transmitted, and it is given a media access control (MAC) address at the point of manufacture. A MAC address is a unique identifier for that piece of hardware: it is 48 bits long and is usually written as hexadecimal in six pairs of digits, for example 00:1B:44:11:3A:B7. MAC addresses are created using the manufacturer code, which identifies the maker, and the serial code, which identifies the individual device. Because it is fixed at manufacture, a MAC address is used to identify a device on a network.

An IP address is allocated by the network rather than built into the hardware, and it is used to identify a device so that data can be routed to it. An IP address can be static, meaning it never changes, which is used for devices such as servers and printers that must always be found at the same address, or dynamic, meaning it is allocated each time the device connects, which allows addresses to be reused. There are two types in use. IPv4 addresses are 32 bits long and are written as four denary numbers separated by dots, such as 192.168.1.10; because there are only about 4.3 billion of them, the supply has run out. IPv6 addresses are 128 bits long and are written as eight groups of hexadecimal digits separated by colons, which provides vastly more addresses.

A router is the device that sends data to a specific destination on a network. It forwards packets along the best route towards their destination, it can assign IP addresses to the devices on the network, and it can connect a local network to the internet.