What you'll learn
Systems architecture is the design of the central processing unit (CPU) and how it works with memory to run programs. In OCR GCSE Computer Science (J277, section 1.1) you need to know the purpose of the CPU, the von Neumann architecture, the job of each component and register inside the CPU, the stages of the fetch–execute cycle, the three factors that affect CPU performance — clock speed, cache size and number of cores — and what an embedded system is. This guide explains each idea, shows how they connect, and works through the types of questions that appear on Paper 1 (Computer Systems).
Most marks in this topic come from precise use of vocabulary. Examiners expect you to say that the MAR holds an address and the MDR holds data or an instruction, not just that they "store things". They also expect explanations of performance to give a reason: more cache helps because the CPU spends less time waiting for slower RAM.
Key terms and definitions
CPU (central processing unit) — The component that processes data and instructions by repeatedly fetching, decoding and executing them.
Von Neumann architecture — A design in which program instructions and data are stored together in the same memory and the CPU has registers such as the MAR, MDR, PC and accumulator.
ALU (arithmetic logic unit) — Performs arithmetic (addition, subtraction) and logical operations (comparisons, AND, OR, NOT) and binary shifts.
CU (control unit) — Decodes instructions and sends control signals to coordinate the other components; manages the fetch–execute cycle.
Cache — Small, very fast memory inside or very close to the CPU that stores frequently used data and instructions.
Register — A tiny, extremely fast storage location inside the CPU.
PC (program counter) — Holds the address of the next instruction to be fetched.
MAR (memory address register) — Holds the address of the memory location currently being read from or written to.
MDR (memory data register) — Holds the data or instruction that has been fetched from memory, or that is about to be written to it.
Accumulator — Holds the results of calculations carried out by the ALU.
Clock speed — The number of fetch–execute cycles the CPU can perform per second, measured in hertz (e.g. 3.2 GHz).
Core — An independent processing unit; each core can fetch, decode and execute its own instructions.
Embedded system — A computer system built into a larger device to perform a dedicated function.
Core concepts
The purpose of the CPU
The CPU's purpose is to process data and instructions so that the computer can run programs. Every program — a game, a browser, the operating system itself — is a long list of simple instructions stored in memory. The CPU works through them one by one, billions of times a second. Nothing a computer does happens without the CPU carrying out instructions.
Von Neumann architecture
In 1945 John von Neumann described a design that almost every general-purpose computer still follows. Its key idea is the stored program concept: instructions and data are held in the same memory, as binary, and the CPU fetches both from there. Because programs are just data in memory, the same machine can run completely different programs simply by loading different instructions.
The von Neumann design includes a control unit, an ALU, registers (including the MAR, MDR, PC and accumulator) and memory connected to the CPU. Knowing the four named registers is essential for OCR.
Components of the CPU
The ALU carries out all calculations and logical decisions: adding two numbers, comparing values to see whether one is larger, performing logical operations, and carrying out binary shifts.
The control unit decodes each instruction to work out what needs to happen, then sends control signals to the other parts of the CPU, memory and input/output devices so that each step happens at the right time.
Cache stores copies of data and instructions that the CPU is likely to need again. Accessing cache is much faster than accessing RAM, so a "cache hit" saves time. Cache is expensive per byte, which is why it is small compared with RAM.
Registers are the fastest storage of all, located inside the CPU itself. Each has a specific job.
The four registers you must know
- The program counter (PC) holds the address of the next instruction. After each fetch, it is incremented so it points to the following instruction.
- The memory address register (MAR) holds an address — the location in memory to be accessed next.
- The memory data register (MDR) holds the contents fetched from, or to be written to, that address.
- The accumulator holds the result of the ALU's most recent calculation.
The MAR and MDR are the pair students most often mix up. A useful memory hook: the MAR is like the address on an envelope; the MDR is the letter inside.
The fetch–execute cycle
OCR calls this the fetch–execute cycle (it includes a decode stage).
Fetch
- The address held in the PC is copied into the MAR.
- The instruction at that address is fetched from memory and copied into the MDR.
- The PC is incremented to point to the next instruction.
Decode 4. The control unit decodes the instruction in the MDR to work out what it means and what data it needs.
Execute 5. The instruction is carried out. This might involve the ALU performing a calculation (with the result in the accumulator), data being fetched from or written to memory (using the MAR and MDR again), or the PC being changed to jump to a different part of the program.
The cycle then repeats. Jumps work by loading a new address into the PC instead of letting it simply increment — that is how selection and iteration (IF statements and loops) are carried out at the hardware level.
Factors affecting CPU performance
Clock speed. The clock sends regular pulses that keep the CPU's operations in step. Each pulse allows the next step of the cycle. A higher clock speed means more instructions executed per second, so programs run faster. Increasing the clock speed (overclocking) produces more heat and can make the system unstable if cooling is inadequate.
Cache size. A larger cache can hold more of the data and instructions the CPU needs, so the CPU goes to RAM less often. Because RAM is much slower than cache, this reduces waiting and improves performance. There is a limit: beyond a certain size, extra cache gives smaller gains, and cache is costly.
Number of cores. Each core can work on a separate instruction at the same time. A quad-core processor can, in theory, do four times as much work as a single core. In practice the gain is smaller because:
- some programs are not written to split their tasks between cores;
- some tasks must happen in sequence, so they cannot run in parallel;
- the cores have to share resources and communicate, which takes time.
A good answer in the exam never just says "more cores make it faster". It explains how (parallel processing) and, for higher marks, the limitation.
Embedded systems
An embedded system is a computer built into a larger device to perform one specific job. Examples include washing machines, microwave ovens, central heating controllers, car engine management systems, digital cameras, traffic lights and smart thermostats.
Embedded systems are typically:
- dedicated to a single task or small set of tasks;
- small, cheap and low-power, because they do not need general-purpose hardware;
- reliable, because their limited software is fully tested;
- often programmed with firmware stored in ROM or flash memory, which is rarely changed.
By contrast, a general-purpose computer such as a laptop can run any program the user installs. Some embedded devices now connect to the internet (smart devices), which brings convenience but also security risks if they are not updated.
Worked examples
Example 1: Registers during the fetch stage
Describe the role of the MAR and the MDR during the fetch stage of the fetch–execute cycle. (4 marks)
Answer: The address of the next instruction is copied from the PC into the MAR (1). The MAR holds the address of the memory location to be accessed (1). The instruction stored at that address is fetched and placed in the MDR (1). The MDR holds the instruction until it is decoded by the control unit (1).
Example 2: Explaining a performance upgrade
A student's computer has a 2.0 GHz dual-core processor. They are offered a 3.0 GHz dual-core processor with the same cache. Explain how this would affect performance. (2 marks)
Answer: The clock speed increases from 2.0 GHz to 3.0 GHz, so the CPU can carry out more fetch–execute cycles per second (1). Programs will therefore execute more instructions per second and should run faster (1).
Example 3: More cores is not always better
Explain why doubling the number of cores does not always double the speed of a computer. (2 marks)
Answer: Some programs are not designed to split their tasks across multiple cores, so the extra cores are not used (1). Some tasks depend on the result of a previous task and must be done in sequence, so they cannot run in parallel (1).
Example 4: Identifying an embedded system
State whether each is an embedded system and justify: (a) a dishwasher controller, (b) a desktop PC.
Answer: (a) Yes — it is built into the dishwasher and has one dedicated function, controlling wash cycles. (b) No — it is a general-purpose computer that can run many different programs.
Common mistakes and how to avoid them
- Swapping the MAR and MDR. MAR = address; MDR = data or instruction.
- Saying the PC "holds the current instruction". It holds the address of the next instruction.
- Saying cache is "on the hard drive" or is "a type of RAM you add". Cache is on or very near the CPU.
- Explaining performance without a reason. Always add because: "more cache means fewer slower RAM accesses".
- Forgetting that clock speed is measured in hertz, not bytes or bits.
- Calling a smartphone an embedded system. It is a general-purpose computer; embedded systems have a dedicated function.
Exam technique for "Systems Architecture"
- For "state" questions, a precise term is enough ("the control unit").
- For "describe" questions about the fetch–execute cycle, give the steps in order using the register names.
- For "explain" questions about performance, give the factor, the effect and the reason. For six-mark extended answers, consider all three factors and their limitations, then reach a conclusion linked to the user's needs.
- Learn one or two everyday examples of embedded systems and be ready to explain why they qualify.
Quick revision summary
- The CPU processes data and instructions by repeating the fetch–execute cycle.
- Von Neumann architecture: instructions and data stored together in memory; registers MAR, MDR, PC and accumulator.
- ALU calculates and compares; CU decodes and coordinates; cache stores frequently used data close to the CPU.
- PC = address of next instruction; MAR = address being accessed; MDR = data/instruction fetched; accumulator = ALU results.
- Fetch: PC → MAR → memory → MDR, PC incremented. Decode by the CU. Execute.
- Performance: clock speed (cycles per second), cache size (fewer RAM accesses), cores (parallel processing, with limits).
- Embedded systems are dedicated computers inside larger devices: small, cheap, low-power, reliable.