問67
頻出度★☆☆
CPUに関する記述のうち, 適切なものはどれか。
平成23年度 特別試験 ITパスポート試験 問67
テクノロジ系コンピュータ構成要素 › プロセッサ
This question's explanation and your answers so far.
解説
問 67
Incorrectア
This description confuses cache memory with virtual memory, so it is wrong. Cache memory is a high-speed but small-capacity buffer memory placed between the CPU and main memory in order to bridge the gap in operating speed between them. It holds data that has been read or written once together with data near it, so that the next time the same area is needed there is no need to go all the way to the slow main memory; this shortens the effective access time and speeds up processing. Its purpose is solely to improve speed, not to increase the capacity of main memory. The mechanism that makes main memory look larger than it is, on the other hand, is called virtual memory: it uses part of auxiliary storage such as a hard disk in place of main memory and swaps only the parts that are needed in and out of main memory (paging), making it possible to run programs larger than the actual main memory capacity. It helps to remember them as a pair: cache makes up for speed, virtual memory makes up for capacity.
Incorrectイ
This choice states the relationship between the arithmetic/logic function and the control function incorrectly, so it is wrong. The five major units of a computer are the input unit, the output unit, the storage unit, the arithmetic/logic unit, and the control unit; of these, the arithmetic/logic unit and the control unit taken together make up the CPU (central processing unit). The arithmetic/logic unit (ALU) handles the actual calculation work: arithmetic operations such as addition, subtraction, multiplication, and division, logical operations such as AND, OR, and NOT, and magnitude comparisons. The control unit fetches instructions from main memory and decodes them, and according to their content issues directions to the arithmetic/logic unit, the storage unit, and the input/output units, controlling the order in which the whole machine operates. The two are equal functions that make up the CPU, and neither is contained within the other. The statement “the arithmetic/logic function is a part of the control function” therefore does not hold.
Incorrectウ
The first half, which explains the classification by number of bits, is correct, but the assertion at the end that “the number of bits has no relation to processing capability” is wrong. The number of bits here refers to the width of the data the CPU can handle all at once (the width of its registers and data bus). A 16-bit CPU can handle only 16 bits at a time, while a 32-bit CPU can handle 32 bits at a time, so it needs fewer passes to process the same amount of data and its processing capability is correspondingly higher. In addition, the width of the addresses it can handle at once also becomes wider, which has the effect of increasing the amount of main memory it can use directly. In other words, the number of bits is an important factor governing processing capability, not an unrelated one. Note, however, that processing capability is not determined by the number of bits alone: the clock frequency, the number of cores, the cache capacity, and the internal architecture all have a combined influence as well, and that is worth understanding too.
- エ
正解 The correct answer is エ. A CPU operates in step with a periodic synchronization signal inside it called the clock, and it fetches, decodes, and executes instructions in time with that signal. The clock frequency expresses how many times per second this synchronization signal ticks; its unit is Hz (hertz), so 1 GHz means one billion cycles per second. The time one cycle takes (the clock cycle time) is the reciprocal of the frequency, so the higher the frequency, the shorter the time each tick takes. What deserves attention here is that the choice explicitly states the condition “for CPUs with the same architecture”. Having the same architecture means the instruction set, the number of pipeline stages, and the number of clock cycles needed to execute one instruction (CPI) are all identical, which means the amount of work done per clock cycle is the same. Since the amount of work is the same and only the ticking gets faster, the number of instructions that can be executed per unit of time increases in proportion to the frequency, and the processing speed becomes faster. This description is therefore appropriate. Conversely, once the premise “if the architecture is the same” is removed, the story changes. Between CPUs of different architectures, the number of instructions handled per clock, the number of cores, and the cache capacity all differ, so comparing clock frequency alone does not let you judge which is faster. It really does happen that a CPU with a lower frequency outperforms one with a higher frequency thanks to a more efficient architecture or a larger number of cores. It is worth understanding this point about comparison under a stated premise as well.
ここがポイント
- The relationship between clock frequency and processing speed — the clock frequency is the number of times per second that the synchronization signal setting the CPU's pace ticks (unit: Hz), and the higher it is, the shorter one clock cycle lasts. Between CPUs of the same architecture (the same architecture, the same CPI), the higher the frequency the faster the processing speed; but between CPUs of different architectures the number of cores, the cache, and the amount of work done per clock all differ, so frequency alone cannot tell which is superior.
- The composition of the CPU, cache, and the number of bits — the CPU is made up of the arithmetic/logic unit and the control unit among the five major units, and the two are equal functions with no containment relationship. Cache memory is a high-speed, small-capacity memory that bridges the speed gap between the CPU and main memory (a supplement for speed), while what apparently expands the capacity of main memory is virtual memory, which uses auxiliary storage (a supplement for capacity). Also, the number of bits, which is the width of data that can be handled at one time, is directly connected to processing capability.
用語