What is the role of the CPU clock?
The clock produces timing pulses used to coordinate CPU activity. Clock speed is the number of cycles per second, although more cycles alone do not determine total performance.
Specification H446. Explore a topic, practise the skill, and see where to focus next.
Try three questions →Explore the topic map ↓
OCR · Computer Science A-Level · H446LearningP is independent of OCR · Computer Science A-Level. Use the current official documents for assessment requirements.
Try three different Computer Science skills from this OCR · Computer Science A-Level route. See what went right, understand a mistake, and find a useful next step.
A small preview of how LearningP turns answers into a clearer learning picture for students and parents.
The clock produces timing pulses used to coordinate CPU activity. Clock speed is the number of cycles per second, although more cycles alone do not determine total performance.
A RISC design can use simpler control circuitry and simple instructions, which can reduce energy consumption and make pipelining easier. Suitability still depends on the workload and implementation.
Variable-size allocation can leave scattered gaps that are collectively large enough but not contiguous enough for a request. This is external fragmentation.
These tiles show your answers to three questions. They are a starting point, not a mastery score or grade prediction.
Correct answer: It sends regular timing pulses that synchronise processor operations
The clock produces timing pulses used to coordinate CPU activity. Clock speed is the number of cycles per second, although more cycles alone do not determine total performance.
Correct answer: Simple instructions and simpler circuitry can reduce power use
A RISC design can use simpler control circuitry and simple instructions, which can reduce energy consumption and make pipelining easier. Suitability still depends on the workload and implementation.
Correct answer: Free memory is split into small non-contiguous holes
Variable-size allocation can leave scattered gaps that are collectively large enough but not contiguous enough for a request. This is external fragmentation.
For parents: look at the explanation together. A correct answer is encouraging; a missed answer gives you something specific to work on. Broader practice over time is needed to understand progress.
Modern processors combine several techniques and may use a modified Harvard cache arrangement while presenting a unified address space. Real designs need not match one simplified model exactly.
Watch for: Do not describe Structure and function of the processor vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Only the parallelisable part of a program benefits from extra cores. Sequential work, communication, synchronisation and scheduling overhead can limit the speed-up, so eight cores do not guarantee four times the performance.
Watch for: Do not describe Types of processor vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Virtual storage hides the physical location of stored files or volumes, which may be remote. Virtual memory uses disk or solid-state storage as an extension to RAM and is much slower than physical memory.
Watch for: Do not describe Input, output and storage vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Shortest remaining time continually favours the process with the least estimated work left and can pre-empt the current process. Shortest job first is normally non-pre-emptive.
Watch for: Do not describe Systems software vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Open-source licences grant specified rights to inspect, modify and redistribute source. Closed-source licences generally restrict those rights; cost, quality and support vary in both models.
Watch for: Do not describe Applications generation vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Methodology selection should be justified from project characteristics: requirement stability, technical risk, regulatory needs, team size, timescale and access to stakeholders. No single model is best for every project.
Watch for: Do not describe Software development vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
A class defines attributes and methods; an object is a concrete instance created from that definition. Different objects can hold different attribute values while sharing class behaviour.
Watch for: Do not describe Types of programming language vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Asymmetric encryption uses a public/private key pair. A sender can use the recipient's public key without a prior secret exchange, while only the recipient's private key decrypts the message.
Watch for: Do not describe Compression, encryption and hashing vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Record locking coordinates concurrent transactions so another transaction cannot make an incompatible change at the same time. Locks must be managed carefully to avoid excessive waiting or deadlock.
Watch for: Do not describe Databases vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Client-server designs support central administration and consistent services but depend on server capacity and availability. Peer-to-peer designs distribute resource provision but are harder to manage consistently.
Watch for: Do not describe Networks vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
PageRank does not count every link equally. Rank passed by a source depends on that source's own rank and how many outgoing links share it, so one authoritative link can outweigh many weak links.
Watch for: Do not describe Web technologies vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Hexadecimal is a compact human-readable notation: one hex digit maps exactly to four binary bits. Changing the notation does not reduce the underlying number of stored bits.
Watch for: Do not describe Data types vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
A collision occurs when distinct keys map to the same index. The table must preserve both entries using a collision-resolution strategy such as separate chaining or open addressing.
Watch for: Do not describe Data structures vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
A Karnaugh map arranges truth-table outputs so adjacent cells differ by one input. Groups of powers of two remove changing variables and simplify the expression.
Watch for: Do not describe Boolean algebra vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
A justified analysis considers whether monitoring is authorised, necessary and proportionate, how users are informed, and how collected personal data is secured and retained. Technology alone does not create legal permission.
Watch for: Do not describe Computing-related legislation vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Responsible disclosure gives the organisation an opportunity to reduce harm while avoiding unauthorised access or premature release of exploit details. Exact legal duties depend on jurisdiction and authorisation.
Watch for: Do not describe Moral and ethical issues vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
A weighted graph preserves the airports, available direct flights and fare relevant to the optimisation. Terminal appearance and airline branding do not help find the cheapest route.
Watch for: Do not describe Thinking abstractly vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Reusable functions, classes or modules reduce duplicated code and can improve consistency and maintainability. Their interfaces and assumptions still require documentation and testing.
Watch for: Do not describe Thinking ahead vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Independent parts still need defined interfaces: for example input changes player state, physics updates position, scoring consumes events and display reads the resulting state.
Watch for: Do not describe Thinking procedurally vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
A mutually exclusive selection structure evaluates conditions and follows one matching branch, avoiding execution of several commands for one choice.
Watch for: Do not describe Thinking logically vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Concurrency can improve throughput or responsiveness, but performance depends on available parallel work and correctness depends on synchronisation. Pipelining is a different technique.
Watch for: Do not describe Thinking concurrently vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
A count-controlled loop is appropriate when the number of repetitions is known in advance. A condition-controlled loop is better when repetition depends on a changing condition.
Watch for: Do not describe Programming techniques vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Abstraction makes the problem manageable by representing only properties needed for the computation. The modeller must still check that omitted details do not invalidate the intended route decisions.
Watch for: Do not describe Computational methods vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Binary search compares with the middle item and eliminates half of the remaining sorted list at each step. This normally requires far fewer comparisons than a linear search on a large list.
Watch for: Do not describe Algorithms vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
A project must distinguish computable criteria from subjective judgement. A limited model may be possible if its assumptions and measurable outputs are explicit.
Watch for: Do not describe Problem identification vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
OCR permits groups or personas; a real named end user and formal sign-off are not compulsory. The persona must still be evidence-based and relevant.
Watch for: Do not describe Stakeholders vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Recognising limitations defines honest scope and informs success criteria and future development. It does not weaken a well-justified proposal.
Watch for: Do not describe Research the problem vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Criteria should be both measurable and justified. An impressive but irrelevant threshold does not demonstrate that the solution meets the real need.
Watch for: Do not describe Specify the proposed solution vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Useful decomposition reduces complexity and creates parts that can be designed and tested. One undifferentiated component does not expose solution structure.
Watch for: Do not describe Decompose the problem vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Algorithms must be detailed enough to form a complete solution and should be designed before or within an authentic iterative cycle, not reverse engineered from final code.
Watch for: Do not describe Describe the solution vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Testing during development informs changes; post-development testing evaluates the completed solution. The evidence and purpose should be distinguishable.
Watch for: Do not describe Describe the approach to testing vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Retroactive documentation often misses the decisions and remedial work that demonstrate iterative development. Contemporary evidence is more persuasive and complete.
Watch for: Do not describe Iterative development process vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Repeated halving is generally more efficient on large sorted data, while fitness for purpose also requires evidence that normal, boundary and absent-target cases behave correctly.
Watch for: Do not describe Testing to inform development vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Final robustness testing should attempt to expose failure, not simply confirm the happy path. The moderators describe testing the solution to destruction.
Watch for: Do not describe Testing to inform evaluation vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Development evidence shows how the solution evolved; post-development evidence demonstrates whether the final integrated feature satisfies the criterion.
Watch for: Do not describe Success of the solution vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Implementation alone does not prove usability. A balanced evaluation explains how well the feature supports users and where it falls short.
Watch for: Do not describe Describe the final product vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
Ongoing maintenance includes corrective, adaptive and preventive work. Sensitive data makes patching, recovery and access governance especially important.
Watch for: Do not describe Maintenance and development vaguely. Trace the algorithm, data or system state precisely and test boundary or failure cases.
No matching topics. Try another search.
OCR · Computer Science A-Level · A Level · Computer Science · H446
Version 3.0, April 2026; current checked 28 August 2026
Source checked: 2026-08-28. The current official specification controls assessment requirements and option choices.
Yes. This page offers three original questions from three different skills on this exact course. Each answer has an explanation, followed by a sample heatmap showing what was correct and what to revisit. No account is needed, and taster answers are not saved.
It shows the outcome of these three answers and gives a specific skill to discuss or practise next. It is not a full assessment, a mastery score or a grade prediction. Broader practice over time is needed to understand progress.
LearningP currently maps 37 assessed areas for specification H446. The visible topic map below is derived from the verified route; the current official specification remains controlling.
The verified source bank contains 510 original LearningP question records for this route. Every mapped area meets the current publication minimum and passed the latest blocker and review audit.
No. LearningP uses official specifications and assessment materials to map content and demand, while its practice questions and explanations are independently authored.
No. LearningP reports practice evidence, coverage, strengths and gaps. It does not guarantee or automatically predict examination outcomes.
Use the official OCR · Computer Science A-Level specification and assessment-resource pages linked on this page, together with information supplied by the learner’s school.