What makes one CPU faster than another?
How clock speed, the number of cores, and cache size and type affect the performance of the CPU.
An OCR J277 1.1.2 answer on how clock speed, the number of cores, and cache size and type each affect CPU performance, with worked reasoning on why doubling cores does not double speed.
Reviewed by: AI editorial process; not yet individually human-reviewed
Have a quick question? Jump to the Q&A page
Jump to a section
What this dot point is asking
OCR wants you to explain how three factors change the performance of the CPU: the clock speed, the number of cores, and the size and type of the cache. The marks come from giving the mechanism (why the factor helps), not just naming it, and from knowing the limits, especially why more cores does not simply multiply speed.
Clock speed
The clock speed sets the pace of the cycle, but only helps if there is work ready to do. If the CPU constantly waits for data from slow RAM, a faster clock just means it waits more quickly, which is why cache matters alongside clock speed.
Number of cores
More cores does not simply multiply speed. Two limits apply. First, some programs are sequential, meaning each step depends on the result of the previous one, so they cannot be split across cores and gain little from extra cores. Second, cores share resources such as memory and the bus, so they sometimes have to wait for each other. A program that splits cleanly into independent parts benefits most; a program that does not benefits little.
Cache size and type
How the three factors interact
The three factors improve the cycle in different ways and a balanced processor improves all three. A high clock speed runs more cycles per second, but is held back if the CPU keeps waiting for memory. More cores multiply throughput, but only for work that can be split and only as far as shared resources allow. More or faster cache keeps the needed instructions and data close so fewer cycles are wasted waiting for slow RAM. A processor with a fast clock, several cores and a large cache will usually beat one strong in only a single area, because the strengths cover each other's weaknesses.
Try this
Q1. State what is measured by the clock speed of a CPU. [1 mark]
- Cue. The number of cycles (fetch-decode-execute cycles) the CPU performs per second, in hertz.
Q2. Explain why a quad-core CPU may not run a single program four times as fast as a single-core CPU. [2 marks]
- Cue. Many programs have sequential steps that cannot be split across cores, and cores share resources, so the speed-up is less than four times.
Q3. State what is meant by a cache hit. [1 mark]
- Cue. When the instruction or data the CPU needs is found in the cache rather than having to be fetched from slower RAM.
Exam-style practice questions
Practice questions written in the style of OCR exam questions on this dot point, with worked answer explainers. The year tag is the paper they imitate, not the source.
OCR 20204 marksTwo CPUs have the same clock speed. One has two cores and the other has four cores. Explain how the number of cores can affect performance, and explain why doubling the number of cores does not always double the speed.Show worked answer →
Each core can fetch, decode and execute its own stream of instructions, so a four-core CPU can process up to four instruction streams at the same time, completing more work per second than a two-core CPU when there are several tasks, or one task that has been split into parallel parts.
Doubling the cores does not always double the speed because some programs cannot be split into parts that run independently at the same time (they are sequential, where each step depends on the result of the previous one). Cores may also have to wait for shared resources such as memory or for results from each other.
Markers reward the parallel-processing point and a valid reason performance does not scale perfectly (sequential dependencies or shared resources).
OCR 20223 marksExplain how increasing the size of the cache can improve the performance of a CPU.Show worked answer →
Cache is very fast memory inside or close to the CPU that stores frequently used instructions and data. A larger cache can hold more of the data and instructions the CPU needs, so a greater proportion of requests are found in the cache (a cache hit) rather than having to be fetched from much slower RAM.
This means the CPU spends less time waiting for data, so it can fetch, decode and execute instructions more quickly and complete more work per second.
Markers reward linking cache to "frequently used data", "fewer fetches from slower RAM", and "less waiting / more throughput". Just saying "it is faster" with no mechanism does not earn full marks.
Related dot points
- The purpose of the CPU and the fetch-decode-execute cycle, the von Neumann architecture, and the function of common CPU components (ALU, CU, cache, registers including the MAR, MDR, Program Counter and Accumulator).
An OCR J277 1.1.1 answer on the purpose of the CPU, the fetch-decode-execute cycle, the von Neumann architecture, and the function of the ALU, control unit, cache and the named registers (MAR, MDR, Program Counter, Accumulator).
- The purpose and characteristics of embedded systems, with examples, and how they differ from general-purpose computer systems.
An OCR J277 1.1 answer on embedded systems: what they are, their characteristics (dedicated function, low cost, low power, small size), examples, and how they differ from general-purpose computers.
- The need for primary storage, the purpose and characteristics of RAM and ROM, the differences between them, and the need for virtual memory.
An OCR J277 1.2.1 answer on the need for primary storage, the purpose and characteristics of RAM and ROM, the differences between them, and why virtual memory is needed.
- The need for secondary storage, the common types (optical, magnetic, solid state) and how to choose a suitable device using capacity, speed, portability, durability, reliability and cost.
An OCR J277 1.2.2 answer on the need for secondary storage, the three common types (optical, magnetic, solid state), and how to choose a suitable device by capacity, speed, portability, durability, reliability and cost.
- Why data must be represented in binary, the units of information (bit, nibble, byte, kB, MB, GB, TB, PB) and how to convert between them.
An OCR J277 1.2.3 answer on why computers use binary, the units of information from bit and nibble up to petabyte, and how to convert between units of data capacity.
Sources & how we know this
- OCR GCSE (9-1) Computer Science (J277) specification — OCR (2020)