📚 High-Frequency Exam Topics in IB & CIE Computer Science | IB CIE 计算机：高频考点总结
Mastering computer science requires a deep understanding of core concepts that frequently appear in both IB and CIE examinations. From binary arithmetic to object-oriented design, certain topics are tested year after year, and the marks they carry are remarkably predictable across sessions. This guide consolidates these high-frequency topics, providing bilingual explanations, worked conventions and exam-focused checklists to help students reinforce their knowledge and perform confidently in their assessments.
掌握计算机科学需要深刻理解核心概念，这些概念在 IB 和 CIE 考试中频繁出现。从二进制运算到面向对象设计，某些主题每年都会考察，而且它们所占的分值在不同考季中高度稳定、可预测。本指南汇集这些高频考点，提供中英双语解释、答题规范与应试清单，帮助学生巩固知识，自信应对评估。无论你备考的是 IB Computer Science (HL/SL) 还是 Cambridge IGCSE / AS & A Level Computer Science，下面的十二个板块都是必须反复打磨的核心内容。
1. Number Systems and Data Representation | 数字系统与数据表示
Binary, denary, and hexadecimal conversions are the bedrock of data representation. A nibble (4 bits) maps directly to a single hex digit, making hex a compact way to represent large binary values such as memory addresses and colour codes. Examiners rarely ask for conversions in isolation; they embed them inside questions about bitmaps, sound sampling or instruction addresses.
二进制、十进制和十六进制转换是数据表示的基础。每个半字节（4 位）直接映射到一个十六进制数字，这使得十六进制成为表示长二进制值的紧凑方式，例如内存地址和颜色代码。考官很少单独考查转换，而是把它嵌入到位图、声音采样或指令地址的题目中。
You should also be fluent with units and character encoding. Storage is measured in bits and bytes, with binary prefixes (KiB, MiB, GiB) based on powers of 2 and decimal prefixes (kB, MB, GB) based on powers of 10 - a distinction that appears in almost every paper.
你还需要熟练掌握存储单位与字符编码。存储以位和字节度量，二进制前缀（KiB、MiB、GiB）基于 2 的幂，十进制前缀（kB、MB、GB）基于 10 的幂——这个区别几乎每份试卷都会出现。字符集方面，ASCII 用 7 位表示 128 个字符，扩展 ASCII 用 8 位表示 256 个字符，而 Unicode 使用更多位（常见 UTF-8、UTF-16）以覆盖全球语言文字与符号。
A quick reference conversion table can speed up your exam work immensely. Learn the first sixteen values by heart, then build larger values by grouping four bits at a time.
一张快速参考转换表可以极大加快你在考试中的作答速度。把前十六个数值背熟，然后每次把四位二进制分成一组，就能快速拼出更大的数值。
| Binary | Denary | Hexadecimal |
|---|---|---|
| 0000 | 0 | 0 |
| 0001 | 1 | 1 |
| 0101 | 5 | 5 |
| 1010 | 10 | A |
| 1101 | 13 | D |
| 1111 | 15 | F |
| 1101 0010 | 210 | D2 |
Common data-representation exam tasks include:
数据表示部分最常见的考题类型包括：
- Converting between binary, denary and hexadecimal, including values larger than one byte. | 在二进制、十进制与十六进制之间转换，包括大于一个字节的数值。
- × colour depth, and of sampled sound using sample rate × duration × resolution × channels. | 用 宽 × 高 × 色深 计算位图文件大小，用 采样率 × 时长 × 分辨率 × 声道数 计算声音文件大小。
- Explaining why data loss occurs when reducing colour depth or sample resolution. | 解释降低色深或采样分辨率为何会造成数据丢失（精度下降）。
2. Two's Complement and Binary Arithmetic | 补码与二进制运算
Negative integers are often stored using two's complement. To find the two's complement, invert all bits and add 1. This representation simplifies subtraction by allowing addition of the complemented value, which means the processor needs only one adding circuit for both addition and subtraction.
负整数通常使用补码（two's complement）存储。要得到补码，先将所有位取反，然后加 1。这种表示简化了减法，因为可以用「加上被减数的补码」来实现减法运算，处理器因此只需要一套加法电路即可同时完成加减法。
Two's complement = invert all bits (one's complement) + 1
补码 = 按位取反（反码）+ 1
In an 8-bit two's complement system, the leftmost bit is the sign bit: 0 means positive and 1 means negative. The range of values is therefore -128 to +127. A common exam question gives the denary value -45 and asks for its 8-bit representation, or gives 1101 0011 and asks for the denary value using the place values -128, 64, 32, 16, 8, 4, 2, 1.
在 8 位补码系统中，最左边一位是符号位：0 表示正数，1 表示负数，因此取值范围是 -128 到 +127。常见考法是给出十进制数 -45 要求写出 8 位表示，或者给出 1101 0011 要求按位值 -128、64、32、16、8、4、2、1 计算其十进制值。
Overflow is the error that occurs when the result of a calculation falls outside the range that can be represented with the available number of bits. Adding two positive numbers and obtaining a negative result is the classic symptom. You should also be able to explain logical shifts, where a left shift multiplies by 2 and a right shift divides by 2, with the sign bit preserved in an arithmetic right shift.
溢出（overflow）是指计算结果超出了当前位数所能表示的范围而发生的错误。两个正数相加却得到负数结果，就是最典型的表现。你也应能解释逻辑移位：左移相当于乘以 2，右移相当于除以 2；而算术右移会保留符号位。若题目要求判断是否溢出，最直接的方法是先把二进制转成十进制验算，再说明超出 -128 至 +127 的范围。
- Always state the number of bits you are assuming. | 作答时务必写明你假设的位数。
- Show the inversion and the +1 as two separate steps to secure method marks. | 把「取反」和「加 1」分成两步写，以拿到过程分。
- Finish by converting back to denary to check the sign and magnitude. | 最后转回十进制，核对符号与数值是否正确。
3. Floating-Point Representation and Normalisation | 浮点数表示与规格化
making the first bit after the point a 1, that is, storing it in the form 0.1...) maximises precision, because it removes leading zeros that carry no information.
浮点数表示为 尾数 × 2^指数（例如 0.1101 × 2³）。对尾数进行规格化（使小数点后第一位为 1，即存成 0.1… 的形式）可以最大化精度，因为它去掉了不携带任何信息的无意义前导零。规格化的另一好处是同一数值只有唯一表示，便于比较。
The number of bits allocated to the mantissa determines precision, while the number of bits allocated to the exponent determines range. IB and CIE exams often ask you to convert between binary floating-point and denary, to normalise an unnormalised value, and to identify the effect of increasing the exponent field.
分配给尾数的位数决定精度，分配给指数的位数决定范围（可表示的最大值与最小值）。IB 和 CIE 考试经常要求你在二进制浮点数和十进制数之间进行转换、把非规格化数规格化，以及说明增加指数位数所带来的影响。作答时通常需要分四步：确定符号、写出规格化尾数、计算指数、给出二进制与十进制两种结果。
Exponent bias, where a fixed offset is added so that both positive and negative exponents can be stored as unsigned binary, also appears regularly at A Level. If a paper states "exponent bias of 8" for a 4-bit exponent, remember that the stored value equals the true exponent plus 8.
指数偏置（exponent bias）也常在 A Level 出现：给指数加上一个固定偏移量，就能把正指数和负指数都用无符号二进制存下。如果题目给出「4 位指数、偏置为 8」，要记住存储值 = 真实指数 + 8。与之相关的常见失分点是只写结果、不写规格化过程，或者忘记指数可以取负值。
4. Boolean Logic and Logic Gates | 布尔逻辑与逻辑门
Boolean algebra underpins every digital circuit on the syllabus. You must be able to construct a truth table for a given expression, simplify the expression, and then draw the corresponding circuit using NOT, AND, OR, NAND, NOR and XOR gates.
布尔代数（布尔逻辑）是考纲中所有数字电路的基础。你必须能够为给定表达式构造真值表、化简表达式，然后用 NOT、AND、OR、NAND、NOR 和 XOR 门画出对应的逻辑电路图。题目常从「由文字描述画电路」和「由电路写布尔表达式」两个方向出题。
The identities below account for a large share of the marks in this topic:
下面这些恒等式在本主题中占据了相当大比例的分值：
- De Morgan's laws: NOT (A AND B) = (NOT A) OR (NOT B); NOT (A OR B) = (NOT A) AND (NOT B). | 德摩根定律：NOT (A AND B) = (NOT A) OR (NOT B)；NOT (A OR B) = (NOT A) AND (NOT B)。
- Absorption: A OR (A AND B) = A; A AND (A OR B) = A. | 吸收律：A OR (A AND B) = A；A AND (A OR B) = A。
- Distributive: A AND (B OR C) = (A AND B) OR (A AND C). | 分配律：A AND (B OR C) = (A AND B) OR (A AND C)。
es Sum = A XOR B and Carry = A AND B; a full adder adds a carry-in and can be chained to build a multi-bit adder.
卡诺图（Karnaugh map）提供了图形化的化简方法，半加器与全加器则是最标准的应用题。半加器输出 Sum = A XOR B、Carry = A AND B；全加器多了一个进位输入（carry-in），可以串联构成多位加法器。答题时不要漏掉真值表的每一行，圈组时务必使用 2 的幂大小的矩形分组。
5. CPU Architecture and the Fetch-Execute Cycle | CPU 架构与取指-执行周期
The von Neumann architecture model appears in nearly every specification: a single processor, shared main memory, and data and instructions stored together. Within the CPU, the control unit manages timing and decoding, the ALU performs arithmetic and logic, and registers provide the fastest temporary storage.
冯·诺依曼架构模型几乎出现在每一份考纲中：单一处理器、共享主存、指令与数据存放在同一存储器中。CPU 内部，控制器（Control Unit）负责定时与指令译码，算术逻辑单元（ALU）执行算术与逻辑运算，寄存器则提供最快的临时存储。
Register knowledge is consistently examined:
寄存器的作用是稳定考点：
- PC (Program Counter): holds the address of the next instruction. | PC（程序计数器）：保存下一条要执行指令的地址。
- MAR (Memory Address Register): holds the address currently being read or written. | MAR（存储器地址寄存器）：保存当前正在读写的地址。
- MDR (Memory Data Register): holds the data or instruction just fetched or about to be stored. | MDR（存储器数据寄存器）：保存刚取回或即将写出的数据或指令。
- CIR (Current Instruction Register): holds the instruction being decoded and executed. | CIR（当前指令寄存器）：保存正在译码与执行的指令。
- ACC (Accumulator): stores intermediate results of ALU operations. | ACC（累加器）：保存 ALU 运算的中间结果。
The fetch-execute cycle itself is a four-stage process worth practising as a written answer: the address in the PC is copied to the MAR; the instruction is fetched along the data bus into the MDR and then into the CIR, while the PC increments; the control unit decodes the instruction; and finally the instruction is executed, possibly writing a result back to memory.
取指-执行周期本身是一个四阶段过程，值得作为文字答案反复练习：先把 PC 中的地址复制到 MAR；指令经数据总线取入 MDR，再送入 CIR，同时 PC 自增；控制单元对指令译码；最后执行指令，必要时把结果写回内存。作答时按顺序描述每一步并说明涉及的寄存器与总线，比笼统写「CPU 从内存取指令」得分高得多。
Related high-frequency questions include the role of the address bus, data bus and control bus, the difference between RISC and CISC instruction sets, pipelining, and the effect of cache size on performance.
相关高频题目还包括：地址总线、数据总线与控制总线各自的作用；RISC 与 CISC 指令集的区别；流水线（pipelining）如何提高吞吐量；以及缓存大小对性能的影响。比较类题目通常要求「各写两点并给出场景」，例如 CISC 指令多而复杂、适合编译器生成的代码，RISC 指令少而固定、便于流水线执行。
6. Data Structures: Arrays, Stacks, Queues and Linked Lists | 数据结构：数组、栈、队列与链表
Data structures questions almost always pair a definition with a scenario and a diagram. You should be able to describe static and dynamic structures, explain how each is stored in memory, and justify which one suits a given application.
数据结构题目通常把「定义 + 场景 + 图示」三者结合。你需要能描述静态结构与动态结构、解释它们在内存中的存储方式，并说明哪一种更适合给定应用场景。只写定义而不结合场景，通常只能拿到一半分值。
| Structure | Principle | Typical Use |
|---|---|---|
| Array | Fixed size, direct index access | Lookup tables, pixel data |
| Stack | LIFO, push and pop at top | Undo, recursion, interrupt handling |
| Queue | FIFO, enqueue at rear | Print spooling, keyboard buffer |
| Linked list | Nodes with pointers | Dynamic lists, hash buckets |
A linked list stores each item in a node containing the data and a pointer to the next node. Insertion and deletion are efficient because only pointers change, but access is sequential, so searching is slower than in a static array where the address of any element can be calculated directly.
链表把每一项存放在一个节点中，节点包含数据与指向下一节点的指针。插入和删除效率高，因为只需修改指针；但访问是顺序的，查找比静态数组慢，而静态数组可以通过下标直接计算任意元素的地址。答题时用「内存位置 + 指针位置」两栏画图，往往比纯文字描述更清晰。
Binary search trees, hash tables and graphs with adjacency matrices or adjacency lists also appear, usually coupled with a traversal question: in-order, pre-order and post-order for trees, depth-first and breadth-first for graphs.
二叉搜索树、哈希表和用邻接矩阵或邻接表表示的图也会出现，通常与遍历问题配套：树的中序、前序、后序遍历，以及图的深度优先与广度优先遍历。要能写出给定树的遍历序列，并说明遍历顺序如何用于表达式求值或排序。
7. Algorithms, Searching, Sorting and Complexity | 算法、查找、排序与复杂度
Algorithm questions reward precision. A written answer should state inputs, initialisation, the loop condition, the comparison performed, the termination condition and the output. Flowcharts, pseudocode and program code are all acceptable answer forms, but they must be consistent with each other.
算法题讲究精确。文字答案应写明输入、初始化、循环条件、比较内容、终止条件与输出。流程图、伪代码和程序代码都可以作为答案形式，但三者必须前后一致，例如伪代码里用了数组下标，流程图里就要体现下标变量的变化。
Bubble sort repeatedly compares adjacent items and swaps them, placing the largest remaining value in position at the end of each pass. Insertion sort builds a sorted sublist one item at a time, which makes it efficient for small or nearly sorted data sets. Merge sort divides the list into single-element sublists, then repeatedly merges them in order.
冒泡排序反复比较相邻元素并交换，每趟结束后把剩余最大值放到末位。插入排序逐个构造有序子表，因此对小规模或已接近有序的数据效率较高。归并排序把列表不断二分到只剩单个元素，再有序地两两合并，时间复杂度稳定为 O(n log n)，代价是需要额外的存储空间。
and that each comparison halves the search space - which is exactly why the number of comparisons for 1 000 items is about 10, since 2¹⁰ = 1024.
查找题要区分线性查找 O(n) 与二分查找 O(log n)。要记得二分查找的前提是列表已排序；每次比较都会把查找范围减半——这正是 1 000 个元素约需 10 次比较的原因，因为 2¹⁰ = 1024。题目若要求比较两种算法的适用性，应从数据规模、是否已排序、是否需要频繁插入三方面作答。
8. Programming Paradigms and Object-Oriented Design | 编程范式与面向对象设计
Both IB and CIE require more than one programming paradigm, and both examine object-oriented concepts in detail. A class is a template defining attributes and methods; an object is an instance of that class created at run time.
IB 和 CIE 都要求掌握不止一种编程范式，也都详细考查面向对象概念。类是定义属性（属性/字段）与方法的模板；对象是运行时由该类创建的实例。答题时要能区分「类」与「对象」，并能用具体例子说明二者的关系。
The four pillars of object-oriented design must be defined and applied:
面向对象设计的四大支柱必须既能定义、又能应用：
- Encapsulation: attributes are hidden behind methods, so internal state can only change through a controlled interface. | 封装：属性被隐藏在方法之后，内部状态只能通过受控接口改变，从而提高安全性与可维护性。
- Inheritance: a subclass reuses and extends the attributes and methods of a superclass, reducing duplication. | 继承：子类复用并扩展父类的属性与方法，减少重复代码。
- Polymorphism: the same method call behaves differently depending on the object's class, often through method overriding. | 多态：同一方法调用因对象所属类不同而表现不同，通常通过方法重写（overriding）实现。
- Abstraction: unnecessary detail is hidden so that users and programmers work with a simpler model. | 抽象：隐藏不必要的细节，使使用者与程序员面对更简单的模型。
You should be able to read a UML class diagram and translate it into code, then write a constructor, accessors and mutators. IB also examines recursion: a recursive method must have a base case and a general case that moves towards it, otherwise a stack overflow occurs.
你需要能读懂 UML 类图并将其转换为代码，然后写出构造函数、访问器（getter）与修改器（setter）。IB 还考查递归：递归方法必须有一个基例（base case）和一个不断逼近基例的一般情形，否则会造成栈溢出（stack overflow）。分析递归代码时，最好手工跟踪前三层调用，写清每次调用的参数值。
9. Databases, SQL and Normalisation | 数据库、SQL 与规范化
Relational database questions are highly structured and therefore easy to score in. Learn the vocabulary precisely: a table (relation) stores records (rows) made of fields (columns); a primary key uniquely identifies each record; a foreign key is a primary key used in another table to create a relationship.
关系型数据库题目结构性强，因此是容易拿分的板块。要把术语记准：表（关系）存储由字段（列）组成的记录（行）；主键唯一标识每条记录；外键是出现在另一张表中、用于建立关系的主键。作答时最好画出示意图，用箭头标出一对多关系中的连接字段。
SQL must be written accurately, with correct capitalisation of keywords is not required but consistent syntax is:
SQL 必须写准确，关键字大小写虽不强制，但语法必须一致：
SELECT Field1, Field2 FROM Table WHERE Condition ORDER BY Field1 DESC;
SELECT 字段1, 字段2 FROM 表 WHERE 条件 ORDER BY 字段1 DESC;
Aggregate functions such as COUNT, SUM, AVG, MAX and MIN frequently appear together with GROUP BY, and questions often ask you to explain why normalisation to third normal form reduces data redundancy and prevents update anomalies
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图