The microprocessor can complete instruction fetchingExecute instruction, and with the outside worldstorageandLogic unitExchange information and other operations, yesMicrocomputerThe operation control part of.It can be connected withstorageAnd peripheral circuit chipsMicrocomputer。
Microprocessors and traditionala central processorIn contrast, it has the advantages of small size, light weight and easy modularization.The basic components of microprocessor include:registerHeapArithmetic unit, timing control circuit, and data andAddress bus。
Since the invention of human beings in 1947transistorOver 50 yearssemiconductor technology Experienced silicon transistorIntegrated circuit, Superlarge-scale integrated circuit、VLSIWait several generations,Development speedIt is faster than other industries.Semiconductor technology has a wide impact on the whole society, so it is called "the seed of industry".a central processorIt refers to data processing andProcessingWith the rapid development of large-scale integrated circuit technology, the chip integration density of control components is getting higher and higher,CPUCan be integrated into aSemiconductor chipThis kind of large-scale integrated circuit device with the function of central processing unit is collectively referred to as "microprocessor".It should be noted that the microprocessor itself is not equal toMicrocomputerIt is only the central processing unit of a microcomputer.
Microprocessors are everywhere, whethervideo recorder, intelligent washing machines, mobile phones and other household appliances, or automobile engine control, andnumerical control machine, Missileprecision guidance All kinds of microprocessors should be embedded.Microprocessors are not onlyMicrocomputerThe core component ofIntelligent deviceOfKey components。International super high speedSupercomputer、MainframeHigh end computing systems, such as, are also built with a large number of general-purpose high-performance microprocessors.[1]
internal structure
Announce
edit
microprocessor
16 bit microprocessor (shown in the figure8086 microprocessor)It can be divided into two parts, one isExecutive part(EU), that is, the part that executes the instruction;The other part isBus interface unit(BIU), contact 8086 bus, execute slavestorageThe operation of fetching instructions.After the microprocessor is divided into EU and BIU, the command andExecute instructionsThe operations of are overlapped.There is one EU partregisterHeap, consisting of eight 16 bit registers, can be used to store data, index andStack pointer, Arithmetic operationLogical unit(ALU)ExecutionArithmetic operationandLogical operation,Flag registerConditions for storing the results of these operations.These components in the execution components aredata busTransmitting data.Bus interface unitThere is also oneregisterHeap, where CSDS, SS andESyesstorage space Segment register for segments.IP is the instruction pointer.Internal communicationregisterIt is also a register for temporarily storing data.The instruction queue is taken in advanceInstruction flowStore it.The bus interface unit also has an address adder to segmentregisterValue and offset value are added to obtain 20 bitPhysical address。Data and address viaBuscontrolLogical and8086 outsidesystem bus Contact.8086 has a 16 bit data bus. When the processor and off chip transmit data, 16 bits are transmitted at a timeBinary number。8086 has a primaryPipeline structure, can realize the overlap of on-chip operation and off chip operation.[2]
Classification of microprocessors
Announce
edit
Intel Corporation
According to microprocessorapplication area Microprocessors can be roughly divided into three categories: general-purpose high-performance microprocessorsEmbedded microprocessorandDigital Signal Processor Generally speaking, general-purpose processors pursue high performance, and they are used to runGeneral software, equipped with a complete and complex operating system;Embedded microprocessors emphasize high performance in dealing with specific application problems, and are mainly used to run domain orientedSpecial program, equipped with lightweight operating system, mainly used forCellular telephone、CD playerAnd other consumer appliances;The price of microcontroller is relatively low, and it is in the microprocessor marketrequirementLargest, mainly used for automatic control equipment in automobile, air conditioning, automatic machinery and other fields.
CPU isCentral Processing Unit(Central microprocessor)It is the most important part of a computer and consists of an arithmetic unit and a controller.If computers are compared to people, then CPU is the human brain.The development of CPU is very rapid. It only took 21 years for personal computers to develop from 8088 (XT) to the Pentium 4 era.
development history
Announce
edit
microprocessor
CPUIt has developed for many years. During this periodprocess informationCPU can be divided into: 4-bit microprocessor8-bit microprocessor16 bit microprocessor, 32-bit microprocessor and the latest 64 bit microprocessorpersonal computerThe development of is advancing with the development of CPU.Microcomputer refers to large-scalevlsiIt is the main component to integrate the main components of the computer - controller andArithmetic unitAfter more than 30 years of development, the development of microprocessor can be roughly divided into:
first generation
Phase I
(1971-1973) Generally, 4-bit or 8-bit microprocessors with word length, typically in the United StatesIntel 4004andIntel 8008microprocessor.Intel 4004 is a 4-bit microprocessor that can perform 4-bit binary parallel operations. It has 45 instructions and a speed of 0.05MIPs (Million Instruction Per Second,Millions of instructions per second)。Intel 4004 has limited functions and is mainly used for calculatorsElectric typewriter, CameraPlatform scale, TV sets and other household appliances, so that these electrical equipment have intelligence, so as to improve their performance.Intel 8008 is the first 8-bit microprocessor in the world.storagePMOS process is adopted.At this stage, the computer works slowly, and the microprocessorCommand systemIncomplete,storageVery small capacity, only a few hundred bytes, no operating system, onlyassembly language。Mainly used forIndustrial instrumentProcess control.
Second generation
(1974-1977) Typical microprocessors areIntel 8080/8085,ZilogCorporateZ80And Motorola M6800.Compared with the first generation microprocessor,Integration1~4 times higher,Operation speedIt has been improved by 10~15 times, and the instruction system is relatively complete, with typicalComputer ArchitectureAnd interrupt, directstorageAccess and other functions.
Because microprocessors can be used to complete many computing tasks that used to be done with larger devices, and the price is cheap, semiconductor companies began to compete to produce microprocessor chips.Zilog has produced 8080 enhanced Z80,MotorolaThe company produced 6800,IntelIn 1976, the company produced the enhanced 8085, but these chips basically did not change the basic characteristics of the 8080, and they are all second-generation microprocessors.They all useNMOSProcess, integration degree about 9000transistorThe average instruction execution time is 1 μ S~2 μ S, using assembly languageBASICFortran programming, usingSingle user operating system。
Third generation
The third stage (1978-1984) was 16 bit microprocessor.In 1978, Intel Corporation took the lead in introducing the 16 bit microprocessor 8086. At the same time, in order to facilitate the original 8-bit computer users, Intel Corporation also proposed a quasi 16 bit microprocessor 8088.
8086 microprocessorThe maximum main frequency speed is 8MHz, with 16 bitsData channel,memory addressing The capacity is 1MB.meanwhileIntelIt also producesMathematical coprocessori8087,The two chips use compatibleInstruction set, but the i8087 instruction set has added some speciallogarithm, Index andtrigonometric functionAnd other mathematical calculation instructions.These instruction sets are collectively calledx86Instruction set.Although Intel has produced more advanced and faster new CPUs such as the second and third generation CPUs in succession, they are still compatible with the original x86 instructions, and Intel has followed the original x86 sequence in naming subsequent CPUs untilTrademark registrationIt was a problem that I gave up usingArabic numeralsNaming.
In 1979,IntelThe company has developed 8088.Both 8086 and 8088 use 16 bits in the chipdata transmission, so they are all called 16 bit microprocessors, but 8086 can transmit or receive 16 bit data every cycle, while 8088 only uses 8 bits every cycle.Because most of the original devices and chips are 8 bits, while the external 8 bits of the 8088data transferThe receiver can be compatible with these devices.8088 40 pinDIP package,working frequency For 6.66MHz, 7.16MHz or 8MHz, the microprocessor integrates about 29000 transistors.
After Intel launched 8086 and 8088 CPUs, companies have also launched similar products, including Zilog Z8000 and Motorola M68000.16 bit microprocessors are larger than 8-bit microprocessorsAddressing space, strongerComputing power, Fasterprocessing speed And more perfect instruction system.Therefore, 16 bit microprocessors can replace someMinicomputerEspecially in single task and single user systems, 16 bit microprocessors such as 8086 are widely used.
In 1981, the United StatesIBMThe company will8088 chipIt has been used in the IBM - PC computer it developed, thus creating a new era of microcomputer.It was also from 8088 that the concept of personal computer (PC) began to develop all over the world.Since 8088 was applied to IBM PC, personal computer has really entered people's work and life, and it also marks the beginning of a new era.
In 1982, Intel Corporation developed the 80286 microprocessor based on the 8086 microprocessorDominant frequency20MHz, 16 bit internal and external data transmission, 24 bitMemoryAddressing of the memory,memory addressing The capacity is 16MB.80286 can work in two ways, one is calledReal modeThe other is called protection.
In real mode, the total amount of memory that the microprocessor can access is limited to 1Megabyte;In the protection mode, 80286 canDirect access16 megabytes of memory.In addition, 80286 works in the protection mode, which can protect the operating system, unlike unprotected microprocessors such as real mode or 8086, which will shut down the system in case of abnormal applications.
IBM uses 80286 microprocessor inadvanced techniqueMicrocomputer, namely AT machine, caused a great sensation.80286 has significant improvements over its predecessors in the following four aspects: it supports larger memory;Able to simulateMemory space;Can run multiple tasks at the same time;The processing speed is improved.
The speed of the earliest PC was 4MHz, and the first AT machine based on 80286running speed For 6 MHz to 8 MHz, some manufacturers have also improved the speed by themselves, making 80286 reach 20 MHz, which means that there has been significant progress in performance.
80286 is a kind of packaging called PGAsquarepacking.PGA is derived fromPLCCIt has an internal and external solid pin. In this package, 80286 integrates about 130000 transistors.
IBMPC/ATThe microcomputer bus maintains the three layers of XTBus structure, and added high and low bit busDriverConvert logic and high byte bus.Like XT, the CPU is also welded to the motherboard.
The fourth generation
The fourth stage (1985-1992) is a 32-bit microprocessor.On October 17, 1985, Intel's epoch-making product, the 80386DX, was officially released. It contains 275000 transistors,clock frequency12.5MHz, and then gradually increased to 20MHz, 25MHz, 33MHz, and finally a small number of 40MHz products.
Interior and exterior of 80386DXdata busIs 32-bit,Address busIt is also 32-bit, can address to 4GB memory, and can manage 64TBVirtual storage space。Its operation mode includes real mode andProtection modeIn addition, a "virtual 86" working mode is added, which can be provided by simulating multiple 8086 microprocessors at the same timeMultitaskingAbility.
The 80386DX has more instructions than the 80286. The 80386 with a frequency of 12.5 MHz can execute 6 million instructions per second, 2.2 times faster than the 80286 with a frequency of 16 MHz.The most classic product of 80386 is 80386DX-33MHz, which is generally referred to as 80386.
Due to the powerful computing power of 32-bit microprocessors, PC applications have expanded to many fields, such as business office and computingengineering designAnd computing, data centers, personal entertainment.80386 makes 32-bit CPU the standard of PC industry.
In 1989, Intel Corporation launched a quasi 32-bit microprocessor chip 80386SX.This is for Intel to expandmarket shareA cheaper universal CPU is introduced. Its internaldata bus32-bit, 16 bit external data bus, which can accept 16 bit input developed for 80286/Output interfaceChip, reduce the cost of the whole machine.After the 80386SX was launched, it was widely welcomed in the market because the performance of the 80386SX was much better than that of the 80286, and the price was only one third of that of the 80386.
In 1989, Intel introduced the 80486 chip that we are familiar with.The great thing about this chip, which has been developed for four years and invested 300 million dollars, is that it has broken the limit of 1 million transistors for the first time, integrated 1.2 million transistors, and used 1 micronmanufacturing process。The clock frequency of 80486 is gradually increased from 25MHz to 33MHz, 40MHz and 50MHz.
80486 is a combination of 80386, mathematical coprocessor 80387 and an 8KBCacheIntegrated in one chip.80487 integrated in 80486Operation speedIt is twice as fast as the previous 80387, and the internal cache shortens the microprocessor and slow speedDRAMOfwaiting time。In addition, RISC is adopted for the first time in the 80x86 series(Reduced Instruction Set)Technology can be used in aClock cycleExecutes an instruction within.It also uses burstBusMode, greatly improving the memoryData exchangeSpeed.As a result of these improvements, the performance of the 80486 is four times higher than that of the 80386 DX with the 80387 math coprocessor.
The fifth generation
The fifth stage (1993-2005) isPentiumThe era of (pentium) series microprocessors is usually called the 5th generation.Typical products are Intel's Pentium series chips and compatible onesAMDK6 series microprocessor chip.Internally adoptedSuperscalarinstructionsPipeline structure, withMutual independenceThe instruction and data cache of.along withMMXThe emergence of (MultiMediaeXtend) microprocessor has made the development of microcomputer reach a higher level in networking, multimedia and intelligence.
The early Pentium 75MHz~120MHz used a manufacturing process of 0.5 μ m, and the later Pentium with a frequency above 120MHz used a 0.35 μ m process.The performance of the classic Pentium is fairly average. Integer operations andFloating point operationAll pretty good.In order to improve the application ability of computers in multimedia and 3D graphics, many new instruction sets have emerged. The most famous three are Intel's MMXSSEAnd AMD's 3D NOW!.MMX (Multimedia Extensions) was invented by Intel in 1996Multimedia instructionEnhanced technology, including 57 multimedia instructions, which can process multiple data at one time,MMX technologyWith the cooperation of software, better performance can be obtained.
Multi energy Pentium(PentiumMMX)Official nameThe "Pentium with MMX technology" was released at the end of 1996.Intel has locked and doubled the frequency of its CPUs since the Multi function Pentium. However, MMX's CPU has a strong ability to exceed the external frequencyCore voltageTo super double frequency, so at that timeOverfrequencyIt is a very fashionable action.The word overclocking also became popular at that time.
Multi energy Pentium is another successful product of Intel after Pentium, and its vitality is also quite tenacious.Multi energy Pentium has made significant improvements on the basis of the original Pentium, adding 16KB on the chipData cacheAnd 16KB instruction cache, 4-way write cache andBranch predictionUnit and return stack technology.In particular, the newly added 57 MMX multimedia instructions make the Multifunction Pentium much faster than the Pentium CPU with the same main frequency even when running non MMX optimized programs.
Launched in 1997Pentium IIThe processor combines Intel MMX technology to process movies, sound effects, and drawing materials with high efficiency. Single Edge Contact is used for the first time(S.E.C)Box type package with built-in high-speed cacheMemory。This chip allows computer users to capture, edit, andinternetShare digital photos with relatives and friends, edit and add text, music or make home moviesTransition effect, the use of video phones and the adoption of standardsTelephone lineThe number of Intel Pentium II processor transistors is 7.5 million for transferring movies with the Internet.
Pentium IIIThe processor adds 70 new instructions and Internet streamingSIMDThe extension set, called MMX, can significantly improve the performance of advanced video, 3D, streaming music, movies, voice recognition and other applications. It can significantly improve the use experience of the Internet, so that users can browse realistic online museums and stores, and download high-quality movies. Intel imported 0.25 for the first timeMicron technology,Intel Pentium IIItransistorThe number is about 9.5 million.
In the same year, Intel also released the Pentium III Xeon processor.As the successor of Pentium II Xeon, in addition to adopting a new design on the kernel architecture, it also inherits the new 70 instruction sets of the Pentium III processor to better execute multimediaStreaming MediaApplication software.In addition to facing the enterprise market, Pentium III Xeon has strengthened the ability of e-commerce applications and high-level business computing.In terms of cache speed and system bus structure, there have also been many improvements, which have greatly improved the performanceMultiprocessorCollaborationThe design is carried out.
Launched in 2000Pentium 4With 42 million transistors built in the processor and 0.18 μ m circuits, the speed of the initial release of Pentium 4 is up to 1.5GHzThe number of transistors is about 42 million. In August of the next year, the Pentium 4 processor reached the 2 GHz milestone.In 2002, Intel launched a new Intel Pentium 4 processor with innovative Hyper Threading(HT)Hyper ThreadingTechnology.Hyper threading technology creates a new level of high-performance desktop computer, which can quickly execute multiple computing applications at the same time, or bring higher performance to software supporting multiple threads.Hyper threading technology increases computer performance by 25%.In addition to providing hyper threading technology for desktop computer users, Intel also reached another computer milestone, namely, the introduction of the Pentium 4 processor with an operating frequency of 3.06 GHz, which is the first commercial microprocessor to execute 3 billion computing cycles per second. Such excellent performance is attributed to the most advanced 0.13 micron process technology in the industry at that time. The next year,The Intel Pentium 4 processor with built-in hyper threading technology has a frequency of 3.2 GHz.
PentiumM: ByIsraelNew mobile CPU specially designed by the team,Pentium MyesIntelThe company's x86 architecture microprocessor is used for notebook personal computersBe treated asPart of Centrino, launched in March 2003.The following are publishedDominant frequency: Standard 1.6GHz, 1.5GHz, 1.4GHz, 1.3GHz, low voltage 1.1GHz, ultra-low voltage 900MHz.In order toDominant frequencyGet high efficiency,BaniasIt has been optimized to execute more instructions per clockBranch predictionTo reduce the error prediction rate.In addition, the most prominent improvement is to increase L2 cache to 1MB(P3-M and P4-M are only 512KB). It is estimated that most of the 77 million transistors in Banias are used here.
In addition, there are a series of designs related to reducing power consumption: the enhanced Speedstep technology is indispensable, with multipleSupply voltageAnd computing frequency, so that the performance can better meet the application requirements.
Intelligent power supply distribution can centralize and distribute the system power to the places where the processor needs, and close idle applications;Mobile Voltage Positioning (MVPIV) technology can dynamically reduce voltage based on processor activity, thus supporting lower cooling design power and smallerAppearance design;400 MHz with optimized powersystem bus ;Micro ops fusion technology combines multiple instructions that can be executed at the same time into one instruction to improve performance and power efficiency.Dedicated stack manager, with dedicated hardware to record internal operation, the processor can be nondisruptiveExecution procedure。
BaniasChipset855 series, 855 chipset is composed ofNorth BridgeChip 855 andNanqiao chipICH4-M, the North Bridge chip is divided into 855PM without built-in graphics card (code: Odem) and 855GM with built-in graphics card (code: Montara GM), supporting up to 2GBDDR266/200 memory, AGP4X, USB2.0, two groupsATA-100、AC97Sound effect andModem。The 855GM optimizes the InternalClockGating for 3D and display engines. It can power the 3D display engine when needed, thus reducingChipsetPower of.
Intel launched theDual coreThe processor hasPentium DandPentium Extreme EditionAt the same time, the 945/955/965/975 chipset was launched to support the newly launched dual core processors. The two newly launched dual core processors manufactured with 90nm technology are not usedPinLGA 775 interface, but theSMD capacitorThe number has increased and the arrangement is different.
Desktop platformSandy Bridge Smithfield's processor,Officially namedIt is a Pentium D processor, which is used instead of Arabic numeralsEnglish letterIn addition to representing the generation alternation of dual core processors, the letter D is more likely to remind people of the meaning of dual core processors.
Intel's dual core architecture is more like aDual CPUPlatform, Pentium D processor continues to be usedPrescottArchitecture and 90nmProduction technologyProduction.The Pentium D kernel is actually composed of two independent 2 independent Prescott cores, each of which has an independent 1MBL2 CacheAnd the execution unit. The two cores together have a total of 2MB. However, since the two cores in the processor have independent caches, eachL2 CacheThe information in it is completely consistent, otherwise, an operation error will occur.
To solve this problem, Intel entrusted the coordination between the two cores to the externalMCH(North Bridge)Chips, although the data transmission and storage between caches are not huge, there is no doubt that due to the need for coordinated processing through external MCH chips, the overall processing speed will be delayed to a certain extent, thus affecting the overall performance of the processor.
Since Prescott kernel is adopted, Pentium D also supportsEM64TTechnologyXD bitSecurity technology。It is worth mentioning that Pentium D processors will not support Hyper Threading technology.The reason is obvious:Physical ProcessorAnd multipleLogical processorProper allocation betweendata streamBalancing computing tasks is not easy.For example, ifapplication programTwo computing threads are required. Obviously, each thread corresponds to a physical kernel. But what if there are three computing threads?Therefore, in order to reduce the complexity of dual core Pentium D architecture, Intel decided to cancel the support for Hyper Threading technology in Pentium D for mainstream markets.
The same was made by Intel, and the difference in the names of the Pentium D and Pentium Extreme Edition dual core processors also indicates that the two processors are different in specifications.The biggest difference between them is thatHyper Threading(Hyper Threading) technology.Pentium D NoSupport Hyper Threading TechnologyHowever, the Pentium Extreme Edition does not have such restrictions.With Hyper Threading enabled, the dual core Pentium Extreme Edition processor can simulate the other two logical processors, which can be regarded as four by the systemCore system。
Pentium EESeries are marked with three digits in the form of Pentium EE8xx or 9xx, such as Pentium EE840. The larger the number, the higher the specification or the more features supported.
Pentium EE8x0: indicates that this is Smithfield core, and each core is 1MBL2 CacheThe only difference between the 800MHzFSB and the Pentium D8x0 series is that it adds support for hyper threading technology. In addition, other technical features and parameters are identical.
Pentium EE9x5: indicates that this isPreslerCore, 2MB per coreL2 Cache1066MHzFSB, which is different from the Pentium D9x0 series only by adding support for hyper threading technology andFront end busIt is increased to 1066MHzFSB. In addition, other technical features and parameters are identical.
Single core Pentium 4, Pentium 4 EE, Celeron D and dual core Pentium D and Pentium EE CPUs adoptLGA775Encapsulation.With previousSocket 478The interface CPU is different. The bottom of the LGA 775 interface CPU has no traditional pins. Instead, 775 contacts are used, that is, they are not pin type but contact type. They are contacted with 775 pins in the corresponding LGA 775 slotTransmission signal。LGA 775 interface can not only effectively improve thesignal intensity . Increase the processor frequency, and also improve theYield, Lowerproduction costs。
The sixth generation
Phase 6 (2005 to date) isCoRE The era of (core) series microprocessors is usually called the 6th generation."Core" is a new type of leading energy-savingMicroarchitecture, designed to provide outstanding performance and energy efficiencyPerformance per watt , which is the so-calledEnergy efficiency ratio。The early core is based onNotebook processorOf.Core 2: The English name is Core 2 Duo, which is a new generation of Intel basedCore microarchitectureThe product system of.Issued on July 27, 2006.Core 2 is a cross platform architecture, including server version, desktop version and mobile version.Where, the server versionDevelopment codeWoodcrest, the development code of desktop version is Conroe, and the development code of mobile version isMerom。
The Core microarchitecture of Core 2 processor is designed by Intel's Israeli design teamYonahNew generation based on micro architectureIntel Architecture。The most significant change is to strengthen each key part.To improve the internal data of the two coresExchange efficiencyThe shared two-level cache design is adopted, and the two cores share up to 4MB of two-level cache.
Following the LGA775 interface, Intel first introducedLGA1366Platform, positioning high-end flagship series.First usedLGA 1366The processor code of the interface isBloomfield, improvedNehalemCore, based on 45nm process and original four core design, built in 8-12MBL3 cache。LGA1366 platform introduces Intel Hyper Threading technology again, andQPIBus technologyIt replaces the front-end bus design that has been used since the Pentium 4 era.The most important thing is that the LGA1366 platform supportsThree channel memoryThe designed platform has a greater improvement in actual efficiency, which is also a major difference between the LGA1366 flagship platform and other platforms in terms of positioning.
As the representative of high-end flagship, the processors of early LGA1366 interface mainly include45nmBloomfield CoreCore i7Quad core processor。With Intel's purchase of 32nm process in 2010, the representative of high-end flagship was replaced by Core i7-980X processorcore technology , with the most powerful performance.For users preparing to build high-end platforms, LGA1366 still occupiesHigh end marketThe Core i7-980X and Core i7-950 are still good choices.
IntelCore i7It's a45nmNative quad core processor, the processor has 8MBL3 cache, supportThree channelsDDR3 memory。The processor adopts LGA 1366PinDesigned to support the second generation of hyper threading technology, that is, the processor can run with eight threads.According to the test circulated online, the core i7 on the same frequencyCore 2 QuadThe performance is much higher.
Based on previous data, Intel will first release three Intel Core i7 processors with frequencies of 3.2GHz, 2.93GHz and 2.66GHz,Dominant frequencyThe 3.2GHz processor belongs to Intel Core i7 Extreme, and the price of the processor is $999. Of course, this top processor is aimed at users with fever.The lower frequency 2.66GHz is priced at $284, about 1940 yuanRMB, targeting ordinary consumers.The new generation of Core i7 processors will be available in the fourth quarter of 2013.
From Intel Technology Summit 2008(IDFAccording to Intel's demonstration on 2008), the core i7's capability is about three times that of the core 2 extreme qx9770 (3.2GHz).On IDF,intelThe staff demonstrated the CineBench R10 with a core i7 3.2GHz processorMultithreadingRendering, the result is amazing.After the rendering starts, eight threads of the four cores start working at the same time. Only 19 seconds later, the complete picture appears on the screen, scoring more than 45800.In contrast, core2 extreme qx 9770 3.2GHz can only get around 12000 points,OverfrequencyAt 4.0GHz, it barely exceeds 15000 points, less than one third of the core i7.The super strength of core i7 can be seen from this.
Corei5It is a four core processor based on Nehalem architectureConsolidated memory controller, L3 cache mode, L3 up to 8MB, supportTurbo BoostAnd other technologiesComputer configuration。The main difference between it and Core i7 (Bloomfield) is that the bus does not use QPI, but uses matureDMI(Direct Media Interface) and only supportsDual channelDDR3 memory.Structurally, it usesLGA1156LGA1366 is used in Core i7.I5 has Rui frequency technology, which can be overclocked under certain circumstances.
Core i3Can be seen as a further compact version of Core i5 (orCastration edition), there will be 32nm process version (R&D code:Clarkdale, based onWestmereArchitecture).Corei3The biggest feature is the integration of GPU(Graphics Processor)In other words, Core i3 will be encapsulated by CPU+GPU.Due to the limited performance of the integrated GPU, users can add graphics cards if they want better 3D performance.It is worth noting that even Clarkdale,Show CoreSome fabrication processes will still be 45nm.The biggest difference between i3 and i5 is that i3 does not have Remax technology.
In June 2010, Intel released the revolutionary processor the second generation Core i3/i5/i7 again.The second generation Core i3/i5/i7 belongs to the second generation intelligent core family, all based on the new Sandy Bridge microarchitecture. Compared with the first generation product, it mainly brings five important innovations: 1. The new 32nm Sandy Bridge microarchitecturelow power consumption, better performance.2. Built in high-performance GPU(Core graphics card),video coding , better graphics performance.3、Turbo Boost 2.0, more intelligent and more efficient.4. The introduction of a new ring architecture brings higher bandwidth and lower latency.5. Brand newAVXAES instruction set, strengthen floating point operation and encryption and decryption operation.
SNB(Sandy Bridge)It is a new generation of processor microarchitecture released by Intel in early 2011. The greatest significance of this architecture is to redefine“Integration platform”The concept of "seamless integration" with processors“Core graphics card”It's over“integrated graphics ”The era of.This innovation benefits from the new 32nm manufacturing process.Because the processor under the Sandy Bridge architecture adopts a 32nm manufacturing process that is more advanced than the previous 45nm process, the CPU power consumption has been further reduced in theory, and the circuit size and performance have been significantly optimized, which is to integrate the graphics core (core graphics card) with theCPU packageCreate favorable conditions on the same substrate.In addition, the second generationCoRE New high-definition videoprocessing unit。The high and low speed of video transcoding is directly related to the processorHD videoThe addition of processing unit, the video of the new generation of Core processorprocessing time At least 30% more than the old processor.new generationSandy Bridge ProcessorIt adopts the new LGA1155 interface design, and cannot be compatible without LGA1156 interface.Sandy Bridge is a new micro architecture that will replace Nehalem, but it will still use the 32nm process.What's more attractive is that this time Intel will no longerCPU coreThe GPU core and GPU core are glued together with "glue", but the two are really in the same core.
In the afternoon of April 24, 2012Beijing Planetarium, Intel officially releasedivy bridge(IVB) processor.The 22nm Ivy Bridge will double the number of execution units to a maximum of 24, which will naturally lead to a further leap in performance.Ivy Bridge will support DX11integrated graphics 。Additional XHCI addedUSB 3.0The controller shares four of these channels, providing up to four USB 3.0, thus supporting nativeUSB3.0。The production of the CPU uses 3D transistor technology to reduce the power consumption of the CPU by half.
The arithmetic logic unit ALU mainly completes arithmetic operations (+, -, ×, ÷, comparison) and variousLogical operation(AND, OR, NOTXOR, shifting), etc.ALU YesCombinational circuit, no deposit in itselfOperandsTherefore, two operands must be savedregister:RegisterTMP andaccumulatorAC,The accumulator not only provides ALU with operands, but also receives ALU's operation results.
registerThe array is actually equivalent to the internalRAM, which includesGeneral registerGroup and DedicatedRegister groupTwo parts, general registers (A, B, C, D) are used to store data, intermediate results or addresses involved in the operation.They can be used as two 8-bitregisterTo use.These are inside the processorregisterAfter that, frequent access can be avoidedstorage, can be shortenedInstruction lengthAnd command execution time to improve therunning speed , which also brings convenience to programming.special-purposeregisterincludeProgram counterPC、Stack indicatorSPandFlag registerFR,They are fixed and used to store addresses or address base values.Including:
A)Program counterPC is used to store the address of the next instruction to be executed, so it controls the execution order of the program.Under the condition of sequential execution of instructions, the contents of PC will be automatically increased by 1 for each byte of instructions taken.When the program is transferred, the new instruction address (target address) must be loaded into the PC, which is usuallyBranch instructionTo achieve.
AMD Athlon
B) The stack indicator SP is used to storeTop of stackAddress.stackyesstorageA specific area in.It presses“LIFO”Mode, when new data is pushed instackWhen the data is popped from the stack, the data at the top of the stack will be popped, and the top of the stack will be automatically adjusted.That is to say, the data is being pushedOut of stackThe operation is always carried out at the top of the stack.stackOnce initialized (i.e. determinedBottom of stackPosition in memory), the contents of the SP (that is, the top of the stack position)Automatic management。
C)Flag registerAlso calledProgram status word(PSW)register, used to store the result characteristics after the execution of arithmetic and logic operation instructions, such as carry or overflow flag when the result is 0.
Timing andcontrol logicIt is the core control component of the microprocessor, which is responsible for controlling the whole computer, includingstorageFetch instruction, analyze instruction (i.e. instructiondecoding)Determine command operations andOperandsAddress, fetch operands, execute operations specified by instructions, and send operation results to memory orI/O portEtc.It also sends correspondingcontrol signalTo make the internal and external components of the CPU work in coordination.
It must be pointed out that the microprocessor itself cannot constitute an independent working system, nor can it independentlyExecution procedure, must be equipped with memoryInput/output deviceForm a completeMicrocomputerCan work independently.[3]
storage
MicrocomputerOfstorageIt is used to store programs and data currently in use or frequently used.storageIt can be divided into reading and writing modesRandom access memoryRAM (Random Access Memory) andread-only memoryROM(Read only Memory)。RAM is also called read/writestorage,working processThe CPU can read or write its contents at any time as required.RAM isVolatile memoryThat is, its contents will be lost after power failure, so only temporary programs and data can be stored.The contents of the ROM can only be read but not written. The information stored in the ROM remains unchanged after power failure. It is a non-volatile memory.Therefore, ROM is often used to store programs and data of permanent parts.Such as initial guidance programMonitoring programBasic input and output in the operating systemManagement proceduresBIOS, etc.[3]
I/O interface
I/Ointerface circuit yesMicrocomputerImportant components of.He isMicrocomputerConnect external inputsoutput deviceAnd variouscontrol object And communicate with the outside worldinformation switchingLogic control circuit of.Due to the structure, working speed, signal form anddata formatThey are different, so they cannot be directly attachedsystem bus In order to realize information exchange with CPU, I/O interface circuit must be used for intermediate conversion.I/O interfaceAlso called I/OAdapterDifferent peripherals must be equipped with different I/O adapters.The I/O interface circuit is a microcomputerapplication systemAn essential and important component.In fact, the development and design of any microcomputer application system is mainly the development and design of I/O interface.Therefore, I/O interface technology is one of the important contents discussed in this course, which will be introduced in detail in Chapter 8.[3]
Bus
Busyescomputer systemTransmission of information between components inPublic access, YesMicrocomputerImportant components of.It consists of several communication lines and a variety ofTristate gateDevice composition.MicrocomputerAlways adopt structural formBus structure, that is, the functional components (microprocessorstorage, I/O interface circuit, etc.)Computer system structureThe uniqueness of.useBus structureAfter that, the relationship between the functional components in the system will be transformed into a single bus oriented relationship between the components(Function board/Card) as long asBus standardCan be connected to the system adopting this bus standard, so thatsystem functionIt is easy to expand or update, simple in structure, and greatly improved in reliability.stayMicrocomputer, depending on their location and application,BusIt can be divided into the following four levels, as shown in Figure 1.4.
(1) On chip bus: it is located inside the microprocessor chip, so it is called chipInternal bus。It is used for interconnection and information transmission between ALU and various registers in the microprocessor (as shown in Figure 1.3Internal busIs the on-chip bus).Due to the limitation of chip area and external pin number, the on-chip bus is mostly singleBus structure, which helpsChip integrationandYieldAnd if it is required to accelerate internal dataTransmission speed, dual bus or triple bus structure can also be adopted.
(2)Chip bus: Chip bus is also called component level (chip level) bus orLocal bus。Microcomputer mainboard, single trigger and other plug-in boards and cards (such as various I/O interface boards/cards)SubsystemThe board/card contains various chips such as CPU, RAM, ROM, I/O interface, etc. These chips are also connected byBusTo connectBecause it is conducive to simplifying the structure, reducing the connection, improving the reliability, and facilitating the transmission and control of information.Usually, the bus connecting chips on various boards and cards is called chip bus or component level bus.
Compared with a completeMicrocomputerFor example, various boards/cards are just a subsystem and a part, so the chip bus is also calledLocal busThe bus used to connect the cards of various functional parts of the microcomputer is calledsystem bus 。Local busIs an important concept, which will be discussed in Chapter 7.
system bus The information uploaded includes data information, address informationcontrol information, therefore, the systemBusThe bus contains three different functions, namelydata busDB(Data Bus)、Address busAB (Address Bus) andControl busCB (Control Bus), as shown in Figure 1.2.
data busDB is used to transmit data information.data busIt's two-wayTristateThe form of bus, that is, it can not onlydata transferreachstorageOr I/O interface and other components, and the data of other components can also be transmitted to the CPU.data busThe number of digits isMicrocomputerAn important indicator of, usually consistent with the word length of microprocessors.For example, Intel 8086 microprocessor has a word length of 16 bitsData bus widthIt is also 16 bits.It should be noted that the meaning of data is broad. It can be real data, instruction code orstatus information , sometimes even a control information, so in actual work,data busThe data uploaded is not necessarily just the real data.
Address busAB is specially used to transmit address, because the address can only be transmitted from the CPU to the outsidestorageOr I/O port, so the address bus is always one-way and tri state, which is the same asdata busDifferent.Address busThe number of bits determines that the CPU canDirect addressingOfMemory spaceSize. For example, if the address bus of an 8-bit microcomputer is 16 bits, the maximumAddressing space2 ^ 16=64KB, 16 bitsmicrocomputerThe address bus of is 20 bits, and its addressable space is 2 ^ 20=1MB.In general, ifAddress busIs n bits, thenAddressing spaceIt is 2 ^ n bytes.
Control busCB is used to transmitcontrol signalandSequential signal。control signalSome microprocessors sendstorageAnd I/O interface circuits, such as read/write signals,Chip Select 、Interrupt responseSignal, etc;Other components also feed back to the CPU, such as interrupt application signalReset signal, bus request signal, limit readiness signal, etc.Therefore,Control busThe transmission direction ofcontrol signalHowever, it is generally bidirectional, and the number of control bus bits should be based on the actual situation of the systemControl needsIt depends.actuallyControl busThe specific situation of depends mainly on the CPU.
K5 YesAMDThe company's first independently produced x86CPU, released in 1996.Because K5 encountered problems in developmentTime to marketthanIntelOfPentiumIt's a lot late. In addition, the performance is poor. ThisUnsuccessfulAMD'smarket shareLarge losses.The performance of K5 is very general, integerComputing powerNot as good asCyrix6x86, but still slightly stronger than Pentium,Floating point operationIts capability is far less than that of Pentium, but slightly better than that of Cyrix.In a word, K5 is a product with average strength.The low price of K5 is obviously more attractive to consumers than its performance. Low price is the biggest selling point of this CPU.AMD is naturally not reconciled to Pentium's influence in the CPU market, so they launched K6 in 1997.K6 CPUDesign indexIt is quite high. It has a brand newMMXInstruction and 64KB L1 Cache (ratioPentiumMMX is doubled), the overall performance is better thanPentium MMX, close to the sameDominant frequencyP Ⅱ level.Compared with K5, K6 can process more instructions in parallel and run on higherclock frequencyOn.AMD is very successful in integer operation, and K6 is slightly backward because it needs to use MMX orFloating point operationOfapplication programCompared with the Pentium with the same frequency, it is much worse.
K6
K6 has 32KB dataL1 Cache, 32KB instruction L1 Cache, integrating 8.8 milliontransistor0.35Micron technology, fifth floorCMOS, C4 processReverse mountingChip, core area 168square millimetre(68 mm2 for new products), using Socket7 architecture.Cyrix is one of themOld qualificationOf CPU developers, as early asx86In the new era, it, Intel and AMD have formed a situation of three powers standing side by side.
K6-2
AMD officially launched K6-2 microprocessor in April 1998.It is manufactured by 0.25 micron process, the chip area is reduced to 68 square millimeters, and the number of transistors is also increased to 9.3 million.In addition, K6-2 has 64KB L1 Cache,L2 CacheIntegrated on the motherboard, the capacity ranges from 512KB to 2MB, and the speed isSystem bus frequencySynchronization,working voltage 2.2V, supportedSocket 7framework.
K6-2 is a K6 chip plus 100MHzBus frequencyAnd support3D Now!Of floating-point instructions“Conjugates”。3D Now!Technology is a major breakthrough in the x86 system, which greatly enhances the density required for processing 3D graphics and multimediaFloating point operationPerformance.In addition, K6-2 supportsSuperscalarMMX technology, supporting 100MHzBus frequency, which means that the system isL2 CacheAnd memoryTransmission rateImprove by nearly 50%, thus greatly improving the performance of the whole system.As the last microprocessor independently developed by Cyrix, Cyrix M Ⅱ began production in March 1998.In addition to the features of 6x86 itself, the microprocessor also supports MMX instructionsCore voltage2.9V, with 256 byte instructions;3.5X frequency multiplication;6.5 million transistors are integrated in the core, with a power consumption of 20.6 watts;64KBL1 cache。
K6-Ⅲ
AMD launched K6 - Ⅲ, code named "Sharptooth" in February 1999, which is the last model of the company to support Super 7 architecture and CPGAPackaging formCPU of 0.25 μ mmanufacturing processThe core area is 135 square millimeters, with 21.3 million transistors integrated, and the operating voltage is 2.2V/2.4V.
Athlon(K7)
Compared with K6-2, the biggest change of K6 - Ⅲ is that 256KB L2 cache (newceleron Only 128KB), andDominant frequencySpeed operation.This change of K6 - Ⅲ will be able to play a greater roleDominant frequencyAdvantages.In addition, the microprocessor also has 64KBL1 cache(32KB for instructions and 32KB for data)System bus frequencySynchronous operationOfL3 cacheIts capacity is between 512KB and 2MB.On June 23, 1999, AMD launched the K7 microprocessor with great strategic significance, andOfficially namedbyAthlon。K7 has two specifications: the first one is manufactured by 0.25 micron process, uses K7 core, and the working voltage is 1.6V (its buffer isDominant frequencyHalf of the speed);The second one is made by 0.18 micron process and uses K75 core;The working voltage is 1.7V and 1.8V.These two types of K7 microprocessors are integrated with 21.3 million transistors,fsb All are 200MHz.
Athlon contains 128 KB L1 Cache (P Ⅱ/P Ⅲ is only 32 KB);512KB~1MB L2 Cache off chip cache.At the same time, it also adopts a new macro processing structure, with three parallel x86 instruction decoders, which can dynamically predict the timing,Out of order execution;K7 has a strongFloating point processing unit, at 3DNOW!With the help of commands, it will have further 3D and multimedia processing capabilitiesFPUMake K7 have twice the performance of other x86 microprocessors!In addition, K7 adopts a new Slot A architecture similar to Slot 1Physical structureThe above two can be interchanged, but the electrical performance of the latter is completely incompatible with the former.stayBusFor the aspect, the Alpha of Digital Company is usedsystem bus Protocol EV6,fsb Up to 200MHz;Athlon is the first AMDSMP(Symmetrical multi microprocessor technology) desktop CPU, that is, users can use Athlon to build dual microprocessor or even four microprocessor systems!AMD launched the new Thunderbird in June 2000(Thunderbirds)、Duron(Duron )Microprocessors once again pose a strong challenge to the microprocessors at the core of Intel Copper.
Thunderbird
Thunderbird is a continuation product of AMD's high-end Athlon series, which adopts a manufacturing process of 0.18 μ m, including Slot A andSocket ATwo different architectures, but they are basically the same in design: both have built-in 128KBL1 cacheAnd 256KB L2 cache, which is the same as CPUDominant frequencySpeed synchronous operation;The working voltage is 1.70V~1.75V, and the corresponding power consumption is smaller than that of the old Athlon;37 million transistors are integrated, and the core area reaches 120 square millimeters.
In addition, Thunderbird microprocessor supports 200MHzSystem bus frequency, provides huge bandwidth, supports Alpha EV6 bus protocol, and has multiple parallel x86 instructionsdecoder。
Duron
Duron microprocessor is AMD's first low-end microprocessor based on Athlon core improvement. Its original R&D code is called "Spitfire".Duronfsb It is also 200MHz, with a built-in 128KBL1 cacheAnd 64KB full speed L2 cache. Its working voltage is 1.5V, so its power consumption is smaller than Thunderbird's.Moreover, its core area is 100 mm2, and the number of internally integrated transistors is 25 million, 3 million more than the Athlon of K7 core.These characteristics are in line with AMD's strategy for low-end market, that is, low costlow power consumptionAnd high performance.In terms of floating point performance, Duron based on K7 system is obviously superior to Intel series microprocessor designed with P6 core. It has three full pipeliningOut of order executionUnit, one for addition/subtraction, one for compound instructions and one for floating pointStorage unit。
Other microprocessor development
Announce
edit
[4]In 1975,IBMThe company has produced several RISC based processors.Among them, 801 is the masterpiece of John Cocke, the father of RISC.Finally, power architecture will be designed 15 years laterSeries of productsA series of chips with RISC structure will emerge several years laterARM
This is a kind of RISC architecture developed after it was recognized by the industry in the 1980s,HPOfHPPA-RISC
In 1975,MotorolaLaunched 6800, with 78 processorsInstruction set。Many types of Motorola monolithic processors and microprocessorsdesign idea They all come from 6800. Even 6809, which was once popular and powerful, inherited 6800 lineage.In 1985, Motorola launched the MC68010 and the 32-bit RISC processor family named 88000.But in 1990, due to the need to developPowerPCAnd was forced to stop production.
Z-80 is designed by Frederico Faggin, who left Intel8-bit microprocessorIt is considered to be the enhanced version of 8080. It is also a very powerful single chip computer at that time. It entered China earlier than the later glorious 51 series. In the early 1980s, the school was based onZ80For basic teaching, this kind of chip is used in the single board computer that needs a TV as a display.
But the firstSingle chipCurrent number of 16 bit processorsTITMS 9900。Although it has a strong momentum since its debut, TI had to reduce the output of 9900 in 1982 in order to develop its DSP business
Semiconductor industryAnother giant,National Semiconductor CorporationLater acquired the X86 series processorCyrixCompany, which was founded by National Semiconductor in 1983(National Semiconductor)The NS32032 is also aRISCProcessor, but unfortunatelyRISCThe architecture's processors arepersonal computerAmong the applications, only POWERPC chips are relatively successful in the market, while others can be said to be very unsuccessful. However, in another area: embedded applications, RISC processors are really brilliant.
In 1981, byStanford UniversityAnd some researchersMIPS。Processor leverages depthPipelining technology。By simplifying the operation cycle of instructions, it solves the bottleneck interlocking problem of the pipeline and facilitatesRISCAn important shift in thinking.
In 1982, byBerkeley College The RISC-I developed has only 32 instructions, pipeline operation and register window, and its performance is superior to the single chip design of the same era
ARMIt is a chip design company that does not produce chips by itself, but throughAuthorized productionTo developARMSeries processors.ARM was founded in a barn in Cambridge, England, in November 1990. At first, there were only 12 people. After more than 11 years of development, today's ARM has more than 700 employees, more than 60% of whom are engaged in research and development,ARMIt is a company that neither produces nor sells chips. It sells chipsTechnical Authorization, establish new microprocessor design, production and salesbusiness model。More importantly, this business model has achieved great success, using ARM technologyIP coreThe microprocessors ofConsumer Electronics, imagingIndustrial control、Mass storage, network, security, wireless and other markets, ARM technology is almost everywhere.ARM licenses its technology to many famous semiconductors, software andOEMEach manufacturer gets a unique set of ARMRelated technologiesAnd services.With this partnership, ARM soon became the founder of many global RISC standards.A total of 30 semiconductor companies have signed hardware technology use agreements with ARMlicense agreement, including Intel, IBMLGLarge companies such as Semiconductor, NEC, SONY, Philips and National Semiconductor.as forsoftware systemPartners, includingMicrosoft, Shengyang andMRIAnd a series of well-known companies.
China R&D
Announce
edit
February 18, 2004, byTsinghua UniversityThe self-developed 32-bit microprocessor THUMP chip finally received the“ID”: Typicalworking frequency 400MHz, power consumption 1.17mW/MHz, 40 chips, maximum operating frequency up to 500MHz, is currently the highest operating frequency microprocessor in China."This marks the beginning of China's independent research and development of CPU chipsSubstantive"The Ministry of Education spoke highly of the birth of THUMP.
At Godson-1Godson-2On the basis of, China is independently developing a new generation of Godson-3.
Godson 3AOfworking frequency 900MHz~1GHz, power consumption about 15W, frequency 1GHzDouble precisionFloating point operationThe peak speed reached 16 billion times per second,Single precisionThe peak speed of floating point operations is 32 billion times per second.Godson 3A adoptsSTCCompany(STMicro) 65 nmCMOS processProduction, 425 million transistors, chipsBGA package, the number of pins is 1121, and the power consumption is less than 15 watts.Godson 3A integrates four 64 bitSuperscalarProcessor core, 4MB two-level cache, two DDR2/3 memory controllers, two high-performanceHyperTransportController, a PCI/PCIX controller, andLPC、SPI、UART, GPIO and other low speed I/O controllers.Godson 3ACommand systemAndMIPS64 compatible and X86 supported by command expansionBinary translation。Godson-3 includes serversHigh performance computer、Low energy consumptionData center, personal high-performance computers, high-end desktop applications, high throughput computing applicationsIndustrial control、digital signal processing , high-end embedded applications and other products have broad market application prospects.