CPU Performance Factors

This lesson covers the three main factors that affect CPU performance: clock speed, cache size, and number of cores. These are specified in OCR J277 Section 1.1.1 and are frequently tested in the exam.

---

Factor 1: Clock Speed

What Is Clock Speed?

The clock speed of a CPU is the number of fetch-decode-execute (FDE) cycles the CPU can complete per second. It is determined by the system clock — an electronic component that generates regular pulses to synchronise the CPU's operations.

How Is It Measured?

Unit	Cycles per second	Example
1 Hz	1 cycle	Original clocks
1 MHz	1 million cycles	Early PCs (1980s)
1 GHz	1 billion cycles	Modern CPUs

A modern desktop CPU might have a clock speed of 3.5 GHz to 5.0 GHz, meaning it can perform 3.5 to 5 billion cycles every second.

How Does Clock Speed Affect Performance?

• A higher clock speed means more FDE cycles per second, so instructions are processed more quickly.
• However, increasing clock speed generates more heat, which requires better cooling.
• There is a practical limit — pushing clock speed too high can cause instability and thermal throttling (the CPU automatically slowing down to prevent damage).

Overclocking

Overclocking means increasing the clock speed beyond the manufacturer's recommended setting. This can improve performance but:

• Generates more heat
• May require additional cooling (e.g. liquid cooling)
• Can reduce the lifespan of the CPU
• May void the warranty

---

Factor 2: Cache Size

What Is Cache?

Cache memory is a small amount of very fast memory located inside or very close to the CPU. It stores copies of frequently accessed data and instructions so the CPU does not have to fetch them from the slower main memory (RAM) every time.

Levels of Cache

Level	Speed	Size	Location
L1 cache	Fastest	Smallest (typically 32-64 KB per core)	Built into each CPU core
L2 cache	Fast	Medium (typically 256 KB - 1 MB per core)	Close to each CPU core
L3 cache	Slower than L1/L2	Largest (typically 4-32 MB)	Shared between all cores

How Does Cache Affect Performance?

• When the CPU needs data, it checks the cache first (cache hit = data found, cache miss = data not found).
• A cache hit is much faster than fetching from RAM because cache is closer to the CPU and uses faster memory technology (SRAM).
• More cache means more data can be stored close to the CPU, increasing the chance of cache hits and reducing wait times.

OCR Exam Tip: If asked why cache improves performance, explain that it reduces the number of times the CPU needs to access slower main memory (RAM).

---

Factor 3: Number of Cores

What Is a Core?

A core is an independent processing unit within the CPU. Each core has its own ALU, CU, and set of registers, and can carry out its own FDE cycle independently.

CPU Type	Number of Cores
Single-core	1
Dual-core	2
Quad-core	4
Octa-core	8

How Do Multiple Cores Affect Performance?

• Multiple cores allow parallel processing — different cores can work on different tasks (or different parts of the same task) at the same time.
• A quad-core CPU can theoretically process four instructions simultaneously.
• However, a quad-core CPU is not four times faster than a single-core CPU because:
  • Not all software is designed to use multiple cores.
  • Some tasks are inherently sequential and cannot be split across cores.
  • There is overhead in coordinating work between cores.

---

Comparing the Three Factors

Factor	What It Does	Advantage	Limitation
Clock speed	More FDE cycles per second	Faster processing of individual instructions	More heat; thermal throttling
Cache size	More data stored close to CPU	Fewer slow RAM accesses	Expensive; diminishing returns
Number of cores	Allows parallel processing	Multiple tasks handled simultaneously	Not all software can use multiple cores

Which Factor Matters Most?

It depends on the workload:

• Gaming / single-threaded applications: Clock speed is often most important.
• Video editing / 3D rendering: Number of cores matters most (these tasks are highly parallelisable).
• General use / office work: A balance of all three factors with adequate cache provides the best experience.

---

Worked Example

Question: A student is choosing between two CPUs for video editing. CPU A has a clock speed of 4.5 GHz and 4 cores. CPU B has a clock speed of 3.2 GHz and 8 cores. Which should they choose?

Answer: CPU B is likely the better choice for video editing. Video editing software is designed to use multiple cores for parallel processing. Although CPU B has a lower clock speed, its 8 cores allow it to process multiple parts of a video simultaneously, which is more beneficial for this type of workload than raw clock speed.

---

Summary

• Clock speed (GHz) determines how many FDE cycles per second (OCR J277 1.1.1).
• Cache stores frequently used data close to the CPU to reduce RAM access times.
• More cores enable parallel processing for multi-threaded workloads.
• Each factor has trade-offs; the best choice depends on the task.

OCR Exam Tip: You may be asked to recommend a CPU for a specific task. Always explain your reasoning by linking the task to the most relevant CPU factor.

---

Worked Example: Comparing Two CPUs for a 3D Rendering Studio

CPU performance questions in OCR J277 almost always require you to weigh clock speed, core count and cache size against the real workload. Walking through a full worked comparison sharpens the habit of linking factor to task.

Scenario: A small 3D animation studio needs to buy a workstation for rendering short films. Rendering is a massively parallelisable workload — each frame (and even each region of a frame) can be processed independently. Artists also need the machine to feel responsive when previewing scenes in real time. Two CPUs are being compared:

• CPU X: 6 cores, clock speed 5.0 GHz, 12 MB L3 cache
• CPU Y: 16 cores, clock speed 3.6 GHz, 48 MB L3 cache

Step 1 — identify the dominant factor for each part of the workload.

• Rendering = parallel -> number of cores dominates.
• Real-time scene preview = usually single-threaded -> clock speed and cache help most.

Step 2 — work through each factor.

Clock speed. CPU X completes 5 billion fetch-decode-execute cycles per second per core; CPU Y completes 3.6 billion. For a single-threaded task, CPU X will finish roughly 5.0 / 3.6 = 1.39 times faster than CPU Y. This means real-time preview will feel snappier on CPU X.

Number of cores. CPU Y has 16 cores to CPU X's 6. In a perfectly parallel rendering workload, CPU Y can process about 16 / 6 = 2.67 times as much work per second as CPU X. After accounting for clock-speed difference (CPU Y is 72% as fast per core), CPU Y still runs ahead on parallel work by about 2.67 x 0.72 = 1.92 — nearly twice as fast on a batch render.

Cache size. CPU Y's 48 MB L3 is four times larger than CPU X's 12 MB. 3D rendering works on large scene data; a bigger cache holds more of that data close to the cores, raising the cache-hit rate and reducing expensive RAM round-trips. Bigger cache compounds the benefit of more cores, because each core independently benefits.