Glossary
Part of the host computer
Operator: execution means of the computer for data processing.
Controller: computer command center, each member of the automatic control by coordinated work.
Memory: a computer storage means, for storing programs and data.
Input device: the machine program and data format information and identification can be received into the computer.
An output device: the result of processing the computer output in some form.
Performance
Machine word: a computer for computing the integer number of bits of binary data can be processed.
Data Path Bandwidth: a data bus that can transfer information in parallel bits.
Main memory capacity: maximum capacity of the main memory can store information.
Throughput: number of systems of processing requests per unit time
Response time: submit a request to the system and to obtain the first time in response to the results produced it needs to be used from the user.
Clock: T period or pulse beat, CPU smallest time unit.
Frequency: the frequency of the reciprocal of the clock period of the internal clock of the main machine.
CPI: the number required to execute an instruction clock cycle
execution time: The time it takes to run a program
MIPS: how many millions of instructions executed per second
FLOPS: how many times to perform floating-point operations per second, respectively MFLOPS, GFLOPS and so on.
Several issues in this section
1) by the computer which of these components?
By the computer 运算器, 控制器, 存储器, 输入设备and 输出设备five major components, modern computer usually 运算器and 控制器integrated on a single chip, collectively known 中央处理器.
In 微处理器Before launch, the arithmetic and control the separation, the memory capacity is small and, therefore designed to be operator-centered structures, other members are passed through the operator information is completed. As follows
With advances in microelectronic technology, and computer processing required, also increasing the amount of information processing, and the CPU speed of a large number of I / O devices disparities speed, so as to center operator can not meet the structural development of the computer. Modern computers have been developed in the memory as the center of the I / O operation to bypass the CPU possible, done directly between I / O devices and memory, to increase the overall efficiency of the system.
2) certain high frequency CPU faster than low frequency CPU it
Indicators to measure CPU speed, there are many, you can not separate a metric to determine the CPU is good or bad. CPU clock speed, i.e., the clock frequency of the CPU core work. It represents the CPU clock speed digital pulse signal shock CPU speed, and the actual frequency of the presence of a certain relationship between the speed of operation, but there is no formula for determining a relationship between the two values can be measured, because the CPU operation speed depends CPU pipeline in all aspects of performance (architecture, cache, instruction set, the number of bits of the CPU, cache size, etc.). As the clock speed does not directly represent the operation speed, and therefore is likely to lower the actual high frequency CPU operation speed phenomenon under certain circumstances.
3) programs written in languages different levels of what is the difference? What language program written in hardware can be directly executed?
Assembly language and machine language instructions corresponding to the machine, but does not directly correspond to the high-level language instruction, having good portability. Machine language which can be directly executed by hardware.
4) the same function may be implemented by software and can be implemented by hardware it?
Hardware and software make two completely different forms, the hardware is solid, the material basis; software is an information, invisible, intangible. But the logic functions, software and hardware are equivalent. Therefore, in the computer system, either directly many of the functions implemented by hardware, and can be implemented in software in cooperation with hardware.
For example, multiplication can with a special multiplier (mainly composed of adders and shifters) implemented, can also be used to achieve multiplication subroutine, and may be implemented by software in cooperation with hardware.
5) Translation differences and connections, assembler, compiler, interpreter?
Translation program refers to the high-level language source code translator took the opportunity Language program (target program) software.
Translation program has two kinds: one is the compiler, which will translate high-level language source code into the target program all at once, each time the execution of the program, simply execute the target program, so long as the source code unchanged, there is no need to re-translation, please Note that the same high-level language in different architectures, compiled into object program is not the same, the target program associated with the architecture, but still lose the program directly executable by the computer hardware. Another explanation is the program that the source of the statement translated into a corresponding machine object code and executed immediately, and then translated to a source language and translation to the next, until all the source language can all be translated and Zhang line
to explain the implementation of the program is the implementation of a translation of an assembler is the kind of language translation program that the assembly language source code into machine language translation program. Steam and do not generate the target program, and to a corresponding machine language. Kinds of machine-oriented low-level language, is symbolic machine language represents
the difference between the compiler and assembler is: If the source language such as C. C ++ Java and other "high-level language", while the Japanese society, such as light with anti-first into machine Beaujolais words. class of "high-quality language" is not in such a rush flat autumn drive parts, machine language and the target language, then such a translation program called assembler.
6) differences and connections between computer architecture and a computer?
Computer architecture is a machine language or assembly language attribute conventional machine programmer visible, comprising a set of instructions, data types, and other memory addressing techniques, mostly belonging to the abstract properties.
Computer Organization refers to how computer architecture embodied in the property, which contains many of the hardware details transparent to the programmer. For example, an instruction system is the structure of the problem, but how to implement the instruction fetch, i.e., analysis command, operand fetch, issue how to belong operation thereof. Thus, when the two machines with the same instruction, only that they have the same structure, as to how the two machine-implemented instructions which, completely different, i.e., they can be considered to be composed of different ways. For example, whether the machine is provided with a multiplication instruction issue a structure, but the realization of the multiplication instruction in what way it is a problem thereof.
7) machine word, instruction word length, word length stored difference and contact?
Machine word: the number of bits of binary data processing computer directly, as equal to the machine word size of the internal registers, which determines the operation accuracy of the computer.
Instruction word length: a number of binary code words contained in the instruction.
Memory word length: length of the binary code stored in a memory unit. They must be an integer multiple of bytes.
Instruction word length stored in an integer multiple of word length and generally, if the instruction word length equal to twice the word length of the memory, it is necessary to remove the secondary fetch an instruction, fetch cycle is twice the machine cycle; if instruction word length equal to the stored word length, it is equal to the fetch cycle machine cycle bursts.
Early computer storage word length instruction word length and the general appearance of the machine and the other data words, so that access - the main storage can be taken out one instruction or data. With the development of the computer, instruction word length variable data word is also variable, but they must all integer multiple of bytes.
Note that 64-bit operating system is a special 64-bit computer architecture designed operating system, it can take advantage of 64-bit processors. However, the machine 64 may be 64-bit operating system, and can use the 32-bit operating system. While the 32-bit processor is unable to use 64-bit operating system.
8) What is transparency? Transparent means what you can see?
In the computer field, standing in certain types of user point of view, if not feel the existence of a thing or property, that is, to "see" the user, not a thing a thing or property, it is called " or property is transparent. " This is the opposite of everyday life "transparent" concept (open and visible).
For example, for a high-level language programmers, floating-point format of these instructions, a multiplication instruction formats, data like how the operation is transparent to the operator: For a machine language or assembly language programmers, instruction format , machine structure, data format is not transparent.
In the CPU, IR, MAR and MDR various types of programmers are transparent.