The CPU and the fetch-execute cycle
Introduit| English | Sens |
|---|---|
| CPU/ˌsiː piː ˈjuː/ | CPU |
| instructions/ɪnˈstrʌkʃnz/ | instructions |
| ALU/ˌeɪ el ˈjuː/ | ALU |
| control unit/kənˈtrəʊl ˈjuːnɪt/ | unité de commande |
| registers/ˈredʒɪstəz/ | registres |
| buses/ˈbʌsɪz/ | bus |
| cache/kæʃ/ | cache |
| embedded system/emˈbedɪd ˈsɪstəm/ | système embarqué |
| clock speed/klɒk spiːd/ | vitesse d'horloge |
The central processing unit
- The CPU 中央处理器 processes data and carries out instructions 指令, repeating a cycle millions of times a second.
- Inside are the ALU 算术逻辑单元, control unit 控制单元 and registers 寄存器, linked by buses 总线.
- It runs every program through the fetch–execute cycle.
Parts of the CPU
- ALU — does calculations and logical comparisons.
- Control unit (CU) — sends control signals to manage all the parts.
- Registers — tiny, very fast stores holding one value each.
- Three buses link them to memory: address (one way), data (both ways), control.


The ALU in a CPU:
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
The arithmetic logic unit does the maths and logic; the control unit manages the parts.
Registers and the fetch–execute cycle
- Special registers: PC (address of the next instruction), MAR (address to access), MDR (data/instruction fetched), CIR (instruction being run), ACC (the ALU's result).
- The cycle repeats three stages: Fetch (copy the instruction into the CPU, then PC + 1) → Decode (the CU works out what it means) → Execute (carry it out).

Le cycle fetch-execute
Tap round the loop the CPU repeats billions of times a second — fetch the instruction, decode what it means, execute it, then do it all again.
Match each CPU part to what it does.
Difficulté 2/5 · 140 pièces + un petit bonus lors de la première réussite aujourd'hui
The ALU computes, the control unit directs, the PC tracks the next instruction, and cache keeps hot data close.
Put the stages of the cycle in order.
Difficulté 3/5 · 200 pièces + un petit bonus lors de la première réussite aujourd'hui
Fetch the instruction, decode what it means, then execute it — repeated continuously.
Cache 高速缓存, performance, embedded systems 嵌入式系统
- Cache is a small amount of very fast memory near the CPU, holding often-used data — so the CPU waits less.
- Performance depends on the number of cores, the cache size, and the clock speed 时钟速度 (cycles per second).
- An embedded system is a small computer built into a device for one fixed job (washing machine, microwave).


Which change would improve CPU performance?
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
More cores, a larger cache and a higher clock speed all improve performance.
Cache is small, very fast memory near the CPU that holds often-used data, so the CPU waits less than it would reading from main memory.
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
A bigger cache, more cores and a higher clock speed all raise performance — cache helps by cutting the wait for data.
Worked example: one trip round the cycle
The program counter (PC) holds 20. Describe how the next instruction is fetched.
- The address 20 is copied from the PC to the MAR and sent along the address bus to memory. The instruction stored at address 20 comes back along the data bus into the MDR, then is copied to the CIR.
- The PC goes up by 1, to 21, ready for the next fetch. The control unit then decodes the instruction in the CIR, and it is executed (the ALU does any calculation, with the result in the accumulator).
Which register holds the address of the next instruction to be fetched?
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
The PC holds the next address and is incremented after each fetch.
Doubling the number of cores always makes every program run twice as fast.
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
Only tasks that can be split between cores speed up, and the cores must share work and memory.
You've got it
- CPU = ALU (maths/logic) + CU (control signals) + registers, linked by address/data/control buses
- key registers: PC (next instruction), MAR/MDR (memory), CIR (current), ACC (result)
- the cycle: fetch → decode → execute
- performance: cores, cache size, clock speed; cache = fast memory near the CPU