Hardware

The central processing unit 中央处理器 (CPU) is the part of the computer that processes data and carries out instructions 指令. It runs programs by repeating a cycle, millions of times each second.

Parts of the CPU

The parts are joined by buses 总线 — sets of wires that carry information:

• the address bus 地址总线 carries memory addresses (one direction);
• the data bus 数据总线 carries data and instructions (both directions);
• the control bus 控制总线 carries control signals.

Special registers

These registers are used during the fetch–execute cycle.

The fetch–execute cycle

The fetch–execute cycle 取指执行周期 has three stages, repeated again and again:

1. Fetch — the instruction is copied from memory 内存 into the CPU (using the PC, MAR and MDR). The PC then increases by 1.
2. Decode — the control unit works out what the instruction means.
3. Execute — the instruction is carried out (the ALU may do a calculation, with the result going to the accumulator).

Cache memory

Cache 高速缓存 is a small amount of very fast memory inside or close to the CPU. It stores data and instructions that the CPU uses often. The CPU can read from cache much faster than from main memory, so the computer runs faster.

What affects CPU performance

Worked example. Two computers are identical except that A has a 3.0 GHz single-core CPU with 2 MB of cache, and B has a 2.4 GHz quad-core CPU with 8 MB of cache. Which is better for video editing, and why? Video editing splits into many tasks that can run at the same time, so B's four cores can process several streams in parallel while A can only work on one at a time. B's larger cache also keeps more frequently used data close to the CPU, so it has to fetch from RAM less often. B is the better choice, even though A has the higher clock speed. Never answer on clock speed alone: weigh cores, cache and clock speed together, and tie each one to the job being done.

Embedded systems

An embedded system 嵌入式系统 is a small computer built inside a larger device to do one fixed job. It is dedicated to that task, is usually small and low-cost, and has no general operating system. Examples: a washing machine, a microwave, a set-top box.

An input device sends data into the computer.

A touch screen can work in three ways:

• resistive 电阻式 — two layers are pressed together where you touch;
• capacitive 电容式 — it senses the tiny electric charge of your finger;
• infrared 红外线 — your finger breaks a grid of light beams.

An output device sends data out of the computer to the user.

Primary storage 主存储器 is memory the CPU can use directly. There are two kinds.

Both are needed: ROM tells the computer how to start, then RAM holds the programs you run.

Secondary storage 辅助存储器 keeps data permanently, even when the power is off. It is non-volatile and is used for long-term storage. There are three types.

When the RAM becomes full, the computer can use part of the secondary storage as extra, pretend RAM. This is called virtual memory 虚拟内存.

Data that is not needed right now is moved out of RAM onto the disk, which frees space in RAM for other programs. This lets you run more programs than the RAM alone could hold. It is slower, because secondary storage is much slower than RAM.

Cloud storage 云存储 keeps your data on remote physical servers, reached over the internet, instead of on your own device.

• Advantages: reach it from anywhere on any device; no local storage hardware to buy or maintain; easy to share files and to back up automatically.
• Disadvantages: needs a working internet connection; an ongoing subscription cost; security and privacy depend on the provider; you rely on that provider staying available and in business.

Network interface card (NIC)

A network interface card 网络接口卡 (NIC) is the hardware that lets a device join a network and send and receive data. Each NIC has a built-in MAC address.

MAC address and IP address

• A MAC address (media access control 介质访问控制 address) is a number that identifies one physical device. The maker sets it, and it does not normally change. It is written in hexadecimal.
• An IP address (internet protocol 网际协议 address) is a number that identifies a device on a network. It is given by the network and can change.

So a MAC address stays with the device, while an IP address depends on the network the device is using.

There are two versions of IP address. IPv4 is 32-bit, written as four denary numbers 0–255 (for example 192.168.0.1); the newer IPv6 is 128-bit, written as eight groups of hexadecimal separated by colons, giving vastly more addresses. An address is also either static 静态 (fixed, set by hand and never changing) or dynamic 动态 (handed out by the router/network each time the device connects, so it can change).

Router

A router 路由器 connects different networks together — for example, your home network to the internet. It reads the IP address in each packet and forwards it towards the right network.

• Learn the fetch–decode–execute cycle and the special registers: the program counter holds the address of the next instruction, the MAR/MDR carry the address/data, and the accumulator holds the ALU's result.
• CPU performance improves with more cores, a larger cache, and a higher clock speed.
• RAM is volatile (it loses its data with no power) and holds programs in use; ROM is non-volatile and read-only (the start-up instructions).
• Match the secondary-storage types: magnetic (HDD, tape), optical (CD/DVD, read by a laser), solid state (SSD, USB, flash — no moving parts).
• A MAC address identifies the physical device and does not normally change; an IP address is given by the network and can change.