Microprocessor
A microprocessor ( microprocessor in English) is a Processeur whose components were sufficiently miniaturized so that the whole of the processor can hold on only one Integrated circuit.
Functionally, the processor is the part of a Ordinateur which carries out the arithmetic and logical operations contained in the programs, the latter compose the part Logiciel of this computer.
Description
Until the beginning of the year 1970, different the electronics components forming a processor could not hold on only one integrated circuit. One was to thus place them on several integrated circuits. In 1971, the company Intel succeeded, for the first time, to place all the transistors which constitute a processor on only one integrated circuit thus giving rise to the microprocessor.This miniaturization allowed:
- to increase the frequencies of operation of the processor, because the distances between the components are reduced;
- to reduce the costs, because only one circuit replaces several of them;
- to create computers much smaller: the Microcomputer S.
The passage on Semi-conducteur of certain components is not without posing some problems.
The main features of a microprocessor are:
- the Instruction set which it can carry out. Here some examples of instructions which a microprocessor can carry out: to add two numbers, to compare two numbers to determine if they are equal, to compare two numbers to determine which is largest, to carry out a Sous-programme. A processor can carry out several dozen different instructions.
- the complexity of its architecture. This complexity is measured by the number of Transistor S contained in the microprocessor. The more the microprocessor will contain transistors, the more it will be able to carry out instructions in one second.
-
the number of Bit S which the processor can treat in an instruction. The first microprocessors could not add with the numbers of more than 4 bits in only one instruction. They were to thus carry out several instructions to add with the numbers of 32 or 64 bits. The current microprocessors (in 2007) can treat numbers on 64 bits in only one instruction.
- the maximum speed of the clock which it can support. The role of the clock is to give rhythm the rate/rhythm of the work of the microprocessor. The more the speed of the clock increases, the more the microprocessor Complète instructions in one second.
All this is theoretical, in practice, according to the architecture of the processor, the number of cycles of clock to supplement an elementary operation can vary from a cycle to several tens per unit of execution (typically one on a traditional processor).
For example, a processor has with 400 M Hz can be faster than another B turning him to 1 G Hz, all depends on their respective architectures.
The combination of the preceding characteristics determines the power of the microprocessor. The power of a microprocessor is expressed in MIPS. In the years 1970, the microprocessors supplemented less than one million instructions a second, the current processors (in 2007) can supplement more than 10 billion instructions a second.
History
The microprocessor was invented by Marcian Hoff (called Ted Hoff) in 1971, whereas it was Engineer at Intel.
In 1990, Gilbert Hyatt asserted the paternity of the microprocessor while being based on a Brevet which it had deposited in 1970. The recognition of the anteriority of the patent of Hyatt would have enabled him to claim Redevance S on all the microprocessors manufactured all over the world. However, the patent of Hyatt was invalidated in 1995 by the American office of the patents on the basis of fact that the microprocessor describes in the patent application had not been carried out and could not besides the being with technology available at the time of the deposit of the patent. It seems that Gilbert Hyatt did not give up and hopes to make re-examine this decision.
The first marketed microprocessor, the November 15th 1971, is the Intel 4004 4 - bits. This microprocessor was initially used to manufacture graphic controllers in text mode. It was followed by the Intel 8008. These processors are the precursors of the Intel 8080, Zilog Z80 and of the future family of Intel X86.
The following table describes the main features of the microprocessors manufactured by Intel and watch fulgurating it evolution of the microprocessors as much in increase in the number of transistors, miniaturization of the circuits and increase in power.
Date : the year of marketing of the microprocessor.
Name : the name of the microprocessor.
Transistors : the number of transistors contained in the microprocessor.
Intricacy of engraving : the diameter (in Micrometers) of the smallest wire connecting two components of the microprocessor. In comparison, the thickness of a human hair is of 100 microns!
Frequency of the clock : the frequency of the clock of the mother chart which gives rhythm the microprocessor. MHz = million cycles a second. GHz = billion cycles a second.
Width of the data : the first number indicates the number of bits on which an operation is made. The second number indicates the number of bits transferred at the same time between the memory and the microprocessor.
MIPS : the number of million instructions supplemented by the microprocessor in one second.
Families of microprocessors
A family of microprocessors is the whole of the models of microprocessors which can carry out same the Instruction set basic. This instruction set is a common minimum and often the most recent processors of a family present new instructions which are not always made profitable by preoccupation with a compatibility with the other processors of the family. For example the current code said x86 is in fact of the code x386 which allows the protection memory and which would not function on processors x286.
There exist several families of microprocessors:
-
the family most known by the general public is the family X86, developed mainly by the companies Intel (manufacturer of the Pentium), AMD (manufacturer of the Athlon), VIA and Transmeta. The first two companies now dominate the market and they manufacture most of the microprocessors for Micro-ordinateur S PC compatible. The Intel company also equips the microprocessors for the microcomputers Macintosh since 2006.
-
the microprocessors PowerPC of IBM and Motorola equipped until in 2006 the microcomputers Macintosh (manufactured by Apple). These microprocessors are also used in the waiters of the series p of IBM and in various embarked systems. For the game consoles, a microprocessor derived from PowerPC (Broadway code name) team the Wii, the GameCube, another derivative in three hearts (named Xenon) team the Xbox 360, as for the Playstation 3, it is equipped with a processor different, the Cell, derived from the POWER4, equipped with a principal heart and eight specific hearts.
-
the microprocessor 6502 of the company MOS Technology was used to manufacture celebrates it APPLE II.
-
the microprocessor Zilog Z80 was largely used in the years 1980 in the design of the first personal microcomputers 8-bits like Radio operator Shack TRS-80, Sinclair ZX80, ZX81, ZX Spectrum, standard MSX, Amstrad CPC and later in the embarked systems.
-
the family 6800 of the company Motorola.
-
the family 68000 of Motorola animated old the Macintosh, the Megadrive, the Atari ST and the Commodore Amiga. Their derivatives (Dragonball, ColdFire) are always used in embarked systems.
Among the less known families of the general public:
-
the family Sparc animates most of the waiters and work stations of Sun Microsystems.
-
the family PA-RISC of HP and VLSI Technology, animates the old waiters and work stations of HP, replaced today by family IA-64
-
the family IA-64 of HP and Intel, brings architecture 64 bits to the waiters and work stations of HP
-
the family MIPS animates the work stations of Silicon Graphics, the game consoles like the PSOne, the Nintendo 64 and of the embarked systems, as well as Routeur S Cisco.
-
the family StrongARM is nowadays used only in the embarked systems, whose many PDA and Smartphone S, it was previously used by Acorn for its Archimede S and RiscPC.
-
the family DEC Alpha animated the computers DEC, taken again by Compaq then by HP which definitively stopped it.
Operation
The microprocessors are given rhythm by a Horloge (fast regular signal, imposing a rate/rhythm on the circuit and, possibly ensuring a synchronization with the other components, such as the memory). In the middle of the years 1980, the microprocessors functioned from 4 to 8 MHz. Running 2004, this speed of clock reaches 4 GHz on commercial models (5 GHz in Laboratoire). The higher the speed of the clock is, the more the microprocessor will be able to carry out at high intervals the basic instructions of the programs. But the increase the speed of clock presents disadvantages: the more quickly the microprocessor turns, the more it consumes electricity, and the more it heats.The current microprocessors are optimized to carry out more than one instruction by Cycle of clock, they are microprocessors with paralleled units of execution. Moreover they are equipped with procedures which “anticipate” the following instructions with the assistance of the statistics.
In the race with the power of the microprocessors, two methods of optimization are in competition:
- the technology of the instruction set simplified (RISC, Reduced Instruction Set Computer), rapid with simple instructions of standardized size, easy to manufacture and which one can assemble the speed of the clock without too many technical difficulties.
- technology called CISC (Complex Instruction Set Computer), whose each complex instruction requires more cycles of clock, but which has in its heart much prewired instructions.
Nevertheless, with the considerable reduction in the size of the electronic chips and the gigantic acceleration of the frequencies of clock, the distinction between RISC and CISC quasi completely disappeared. Where distinct families existed, one observes microprocessors today where a structure interns RISC brings to power while remaining compatible with a use of the type CISC (the INTEL series x86 thus discreetly underwent a transition between a very typical organization initially from a structure CISC . Currently it uses a very fast heart RISC , being based on a system of rearrangement of the code to stolen the ) implemented, partly, thanks to memories hiding place increasingly wide, comprising up to three levels.
Structure of a microprocessor
The central processing unit of a microprocessor includes/understands primarily:- a Arithmetic logic unit (UAL) which carries out the operations;
- of the registers which make it possible the microprocessor to temporarily store data;
- a control unit which orders the whole of the microprocessor according to the instructions of the program.
- the indicating register of state (flags), this register gives the state of the microprocessor constantly, it can only be read;
- the meter of program (PC, Program Counter), it contains the address of the next instruction to be carried out;
- the pointer of pile (SP, Stack Pointer), it is the pointer of a special zone of the memory called pile where are arranged the arguments of the Sous-programme S and the Adresse S of return.
The control unit can also break up:
- the instruction register, memorizes the instruction code to be carried out;
- the decoder decodes this instruction;
- the sequencer carries out the instruction, it is him which orders the whole of the bodies of the microprocessor.
See also: Architecture of the processors
Overclocking
Principle
One can translate overclocking into French by the surfréquençage term. Overclocker a processor (or a graphics card) consists in making function this component at an high speed at the normal speed of operation. One thus includes/understands rather quickly the interest of overclocker his processor or his graphics card especially when it is known that the profit can reach up to 10~20% for the majority of the processors. Core 2 (MT) Duet (and to a lesser extent Core 2 (MT) Quad, because of their high thermal dissipation) are distinguished from the remainder of the processors by their incredible capacities of overclocking: 30% for the vast majority of the C2D without effort, up to +50% by taking time to regulate all the parameters correctly, and more still by using more advanced cooling systems (ventirads enormous coppers some, compressors, or even nitrogenizes liquid to go until doubling the initial frequency of the processor).To increase the frequency of the processor increases in fact the speed of the principal data bus of machine (FSB), and thus accelerates all the components connected on the mother chart. The function NCV-Lock, presents on absolutely all the recent cartesmères, makes it possible to avoid this problem and limit the increase in frequency to the processor and the memory.
Risks and dangers
The overclocking is of course under the responsibility of that which practices it and the author of this article could not be blamed for the possible damage that you would subject to your material. The principal risk of the overclocking is to roast the processor by application of a too important supply voltage (Vcore). In the past, the processors also risked cramer if the temperature became too high, but currently absolutely all the processors cut the system automatically if the temperature ateint a dangerous level. The fact of using the processor has a higher frequency also has an influence over its lifespan, even if it is considered in general that the reduction negligible is compared with the life time of a processor (seldom more than 5 years).
Cooling of the processor
The most important problem of the overclocking is thus the cooling of the processor.- the system most usually used is the ventilator assembled on a radiator (the radiator is a metal plate comprising of the wings which makes it possible to improve the exchanges of temperature between the processor on which it is assembled and the ambient air). The ventilator can also be gone up directly on the processor, but cooling will be less good; sometimes a small metal plate intercalated between the ventilator and the processor helps to dissipate the heat of the processor. The ventilator must be bulkiest possible to also allow a mixing of important air which will also contribute to the ventilation of the case…
- the ventilation is, it, very important because it is the mixing of the air of the case which will make it possible to evacuate heat, that the elements provided to the air, outside. Therefore a case " ordonné" allows to minimize the obstacles with ventilation. Indeed tablecloths of the disks hard, and in particular the tablecloths SCSI, which are very broad, if they are located in front of an element which heats (in front of the processor for example) will harm the circulation of air and are likely to cause an overheating (even for a non-overclocké processor); this problem is made negligible by the use of very fine tablecloths Serial-ATA (from 0,5 to 1 mm).
Manufacture of the microprocessors
The manufacture of a microprocessor is primarily identical to that of any Integrated circuit (see the chapter on this subject). It thus follows a complex process. But the enormous size of the majority of the microprocessors tends to increase the cost of the operation further.Law of Moore, which indicates that the level of integration of the microprocessors doubles every 18 months, also states that the production costs double at the same time as the level of integration.
The manufacture of the microprocessors is regarded today as one of the two factors of increase in the capacity of the manufacturing units (with the constraints related to the manufacture of the memories with great capacity). The smoothness of industrial engraving reaches from now on 45 Nm (http://www.intel.com/pressroom/archive/releases/20060125comp.htm). That fixes the maximum speed of treatment of a microprocessor at the neighborhoods of 4GHZ. By decreasing the intricacy of engraving further, the founders run up against the perturbing rules of the quantum Mécanique which upset the results of calculations.
Functions to be developed
Parallel organization
According to the Operating system, the current trend is the installation of several parallel processors and multiple tasks from where growing importance of the functions of arbitrations between process (for example the hyper threading). Indeed, the scalar super architecture (put in parallel of the tasks in a unit of execution) of the current processors is not enough currently any more with the multi-threading such as it is used.On the other hand, the processors with several hearts require that be studied closely the Répartition tasks between them if one does not want to see observing a deceleration of the operations; it is what one names the affinity of the processor ( processor affinity ).
Safety and hiring
There exist many projects of integration in the middle of the microprocessors of functions aiming at preventing illegal copyings of files (technologies DRM). The consortium Trusted Computing Group, in particular, already created chips making it possible to create a " zone of confiance" within the computing system, using a specific chip of identification. Certain models of computers, like the portables of IBM integrate already such chips. The next generation of this technology will be probably integrated in the main frames of the computers.These technologies are décriées, in particular by partisans of the Free software, for whom they have a potential liberticide. Indeed, combined with a Operating system envisaged for this purpose, for example derived from the project NGSCB of Microsoft, this type of technology allows the Tiers confidence (the person receiving benefits which will check the validity of the components of the system) to reach the contents of the computer remotely, to even prevent the execution of certain operations on this one. It should be noted that Linus Torvalds estimates inappropriate to compare a Logiciel with a Contenu, and approves completely the protection of the contained by their owners, although he considers that of the software is not an good idea for those which practice it.
-
Protection and protection, by Linus Torvalds
Wide memory
Old system of extension of the memory allowing to exceed the limit of 1 Me microprocessor 8086 of the time. This memory was accessible by pages of 64 Kio. One nowadays any more does not use it because of the wide capacities of addressings of the recent processors as because of the extreme slowness of the chance accesses as soon as those require a change of memory page.
Operating systems multiple
Vanderpool/Silvervale: if an operating system of work is infected by a virus coriace, another preferably on another basis like Linux towards Mac OS or Windows could act as guard and disinfect first all while not leaving the possibility to the virus of being propagated.
Anticipation of the problems and management remote
Particularly useful for the waiters.
The problem of the heating
The heating of the microprocessors remains roughly and in spite of the use of techniques of increasingly fine engravings, proportional to the square of their tension with architecture given. With the tension, the frequency, and a coefficient of adjustment, one can calculate the dissipated power :
-
a i686 with 1 GHz (1,7 V), twice faster, consumes 34 typically W, which is not far from the quadruple.
- To 2 GHz Opteron dissipates 107 W and a G5 55 W.
This problem is related to another, that of the thermal dissipation and thus often of the Ventilateur S, sources of noise pollutions not easily compatible with an environment of office. Liquid cooling (with water) is proposed.
-
It is useful besides to sandpaper the top of the processor with an aim of removing the roughness of material for a better transfer of heat towards the radiator. A thermal paste has about the same role besides since it ensures a better conduction by absorbing the asperities.
-
a microprocessor containing many transistors will see its power strongly depending on the temperature. This is due to the increasingly random movements of the electrons according to the rise of the temperature.
-
If the heating does not present an main issue for standard application desktop machine, it poses some for all the portable applications. It is technically easy to supply and cool a fixed computer. For the portable applications, they are two delicate problems. The cellphone, the laptop, numerical camera, the PDA, walkman MP3 have a battery which it is a question of sparing so that the portable unit has a better autonomy. In the same way you will not be able to add a ventilator or to consider a liquid cooling on certain portable applications. This is why certain manufacturers propose solutions " low power" , which consumes less energy but operation often at limited frequencies.
See too
Related articles
- Processor
- List of microprocessors
- Chronology of the microprocessors
- Law of Moore
- Digital processing (microprocessor)
- Microcontrôleur
- emulator in-circuit
- Overclocking
External bonds
- Gallery of microprocessors of 1971 to 1996
- Article on the manufacture of the wafers
| Random links: | Epson | Michael Pitt | Madagascar (film) | Manfred von Ardenne | LaMoure |