CPU Design, Caching & Motherboard Communication

Grade 10 introduced the CPU and motherboard as components. Grade 11 asks how they actually talk to each other and what makes one CPU faster than another.

CPU design — techniques used to increase how much work a CPU can do per second:

Multi-processing: A CPU with more than one physical core, so it can genuinely execute multiple instruction streams in parallel (true parallel processing).
Hyper-threading: A single physical core presents itself to the OS as two logical cores, allowing it to juggle two instruction streams more efficiently — it is not true parallel processing, but it improves throughput by keeping the CPU busier.

Registers, ALU and CU (recap from Grade 10, now with more depth):

Registers are the fastest storage in the whole computer, sitting inside the CPU. The number of bits a register can hold (e.g. 32-bit vs 64-bit) directly affects how much data/how large a memory address the CPU can process in one go — a 64-bit register can address far more RAM and handle bigger numbers per instruction than a 32-bit one.
The ALU performs the actual arithmetic and logical operations.
The CU decodes instructions and coordinates the rest of the CPU and its communication with other components.

Performance concepts:

CPU vs RAM performance: the CPU can process data far faster than RAM can supply it — this speed mismatch is called latency, the delay between requesting data and receiving it.
SRAM (Static RAM) is used for processor cache and registers because it's much faster (though more expensive and lower capacity) than the DRAM used for main system RAM.

Caching — keeping frequently-used data closer to where it's needed to reduce latency:

Cache type	Purpose
Processor cache (L1/L2/L3)	Stores recently/frequently used data and instructions right next to the CPU cores
Disk cache	A portion of RAM (or a drive's own memory) used to speed up repeated reads/writes to secondary storage
Browser cache	Stores website files (images, scripts) locally so a page loads faster on a repeat visit
Proxy/web cache	A shared cache (often on a network) that stores web content so multiple users don't need to re-download the same content

Motherboard function and connecting components:

System Clock — a crystal oscillator that generates regular pulses to synchronise the timing of every operation on the motherboard, measured in Hz (cycles per second).
Overclocking and clock multiplication — increasing a component's clock speed beyond its rated value (overclocking) for more performance, at the cost of more heat and reduced stability/lifespan; clock multiplication lets internal CPU components run at a multiple of the base system clock speed.
Internal Bus / FSB (Front-Side Bus) — the set of physical pathways carrying data between the CPU, RAM and chipset: the data bus carries the actual data, the address bus carries the memory location that data should go to/come from, and the control bus carries command/timing signals. Wider buses and higher speeds mean more data can move per cycle.
External buses — connect the motherboard to peripherals/storage; trade-offs are speed vs throughput. PCI Express (PCIe) is used for high-speed components like graphics cards; SATA connects storage drives; USB connects general peripherals.

Example

Why does a 3.5GHz quad-core CPU sometimes feel slower than a 3.0GHz CPU with a bigger cache? Because clock speed alone doesn't determine performance — cache size reduces how often the CPU has to wait (high latency) for slower RAM, so a well-cached CPU can outperform a nominally 'faster' one on real workloads.

💡 Exam Tip

When asked to 'discuss how components interact', always mention the bus system explicitly — most students only describe the CPU and RAM, but the marks are often in explaining how data physically travels between them.