Skip to topic
    ← Back to course topics

    Types of processor — OCR A-Level Computer Science

    Types of processor explained

    Types of processor covers 3 specification points, including (a) The differences between and uses of CISC… and (b) GPUs and their uses (including those not….

    Read the full explanation

    Open a point below for its explanation and revision guidance.

    (a) The differences between and uses of CISC and RISC processors.

    This statement asks you to explain how CISC and RISC processors differ and why each is used. CISC processors have many complex instructions with variable lengths and several addressing modes, so one instruction can take several cycles. RISC processors have a small set of simple, fixed-length instructions, most completing in one cycle, and rely on pipelining and many registers. Programs need more RISC instructions, but the simpler hardware uses less power, which suits smartphones and embedded systems. CISC remains common in desktop and server x86 processors, where compatibility with existing software matters. Link each feature to a use, such as RISC for battery efficiency.

    (b) GPUs and their uses (including those not related to graphics).

    This statement asks you to explain what a graphics processing unit (GPU) is and what it is used for, including uses unrelated to graphics. A GPU has many relatively simple cores that carry out similar calculations simultaneously. This suits rendering images, where each pixel can be processed largely independently. The same design also helps with machine learning and scientific simulation, provided the problem divides into many similar sub-tasks. A CPU, by contrast, has fewer, more powerful cores suited to sequential and varied instructions. In MCQs, pick the option linking many parallel cores to a suitable task, not one calling a GPU a faster CPU.

    (c) Multicore and Parallel systems.

    This statement covers multicore processors and parallel systems. A multicore processor places two or more processing cores on one chip, so separate tasks or threads can run at the same time. Parallel systems extend this idea by using several processors or computers together, each handling part of a problem. For example, a video encoder might split frames across the four cores of a quad-core processor. Gains depend on the task: independent parts benefit, while steps that need earlier results gain little. Shared memory and coordination between processors can also limit the speed-up achieved. In MCQs, choose the option recognising that work must be divisible for parallelism to help.

    Your focus

    1. Compare CISC and RISC instruction sets using at least three contrasting features.
    2. Explain why RISC programs can need more instructions yet still run efficiently.
    3. Justify a suitable use for each design type with a reason linked to its features.
    Show all 9 objectives
    1. Identify the design features of a GPU that make it suited to parallel workloads.
    2. Select a suitable non-graphics use of a GPU from a given scenario.
    3. Explain why a task must divide into similar sub-tasks to benefit from a GPU.
    4. Define a multicore processor and a parallel system.
    5. Describe how work can be divided across cores in a given example.
    6. Evaluate whether a given task is likely to gain speed-up from parallel processing.

    Types of processor exam tips

    Marking Points
    • Identifies that CISC processors use complex, variable-length instructions with several addressing modes.
    • Identifies that RISC processors use a small set of simple, fixed-length instructions that mostly execute in one cycle.
    • Explains that RISC programs need more instructions but each executes quickly, which suits pipelining.
    • Links RISC to low-power and embedded uses, such as smartphones, by reference to its simpler hardware.
    • Links CISC to backward compatibility in desktop and server processors, such as x86.
    • Identifies a GPU as a processor with many relatively simple cores working in parallel.
    • Links rendering of images, where pixels can be processed largely independently, to GPU design.
    • Recognises a non-graphics use such as machine learning or scientific simulation.
    • States that such non-graphics uses need a problem divisible into many similar sub-tasks.
    • Distinguishes a GPU's many simpler cores from a CPU's fewer, more powerful cores suited to sequential tasks.
    • Defines a multicore processor as one chip containing two or more processing cores.
    • Explains that parallel systems use several processors or computers working on parts of one problem at the same time.
    • Gives an example of dividing work across cores, such as frames split across the cores of a processor.
    • States that speed-up depends on whether the task can be divided into independent parts.
    • Identifies shared memory or coordination between processors as a limit on the gains from parallelism.
    Examiner Tips
    • 💡Write each difference as a contrast pair, such as complex versus simple instructions, so that both designs are covered.
    • 💡Use the word so to connect a feature to its consequence, for example so pipelining is effective.
    • 💡Name a real processor family for each design to support your use point.
    • 💡Read each option and check whether it names both the GPU's parallel design and a suitable use.
    • 💡Eliminate options that describe a GPU as only a faster CPU or as useful only for graphics.
    • 💡When a stem mentions machine learning or simulation, check whether the problem can be split into similar parts before choosing that option.
    • 💡Check whether the scenario describes one chip with several cores or several processors working together.
    • 💡Look for whether the task can be split into independent parts before accepting a speed-up claim.
    • 💡Eliminate options that assume every program automatically runs faster on more cores.
    Common Mistakes
    • Says RISC always runs programs faster than CISC; correct this by stating that performance depends on both instruction count and cycles per instruction.
    • Claims CISC instructions are all single cycle; correct this by noting that complex instructions often take several cycles.
    • Gives uses without linking them to features; correct this by stating which feature makes each design suitable for its use.
    • Describes a GPU as simply a faster CPU; correct by stating that its many simpler cores operate in parallel.
    • Limits GPU use to graphics only; correct by including machine learning or scientific simulation as non-graphics uses.
    • Assumes any task suits a GPU; correct by requiring that the work splits into many similar sub-tasks.
    • Treats extra cores as always doubling speed; correct by noting that steps depending on earlier results gain little from parallelism.
    • Confuses a multicore processor with a parallel system; correct by noting that one chip with several cores is multicore, while parallel systems may use several separate processors.
    • Ignores coordination between processors as a limitation; correct by naming shared memory and the overhead of coordinating parts of a task.