【 Weak current College 】 CPU knowledge detailed solution---Power By 【 China power house network 】

1. What is a CPU
CPU is the English "Central Processing Unit/abbreviation for central processing unit" means a unit and controller functionality of large-scale integrated circuits. CPU by logical operations units, control units and storage modules.
2. CPU history
CPU's origins can be traced in 1971. In that year, was still at the initial stage of the Intel Corporation has launched the world's first microprocessor 4004. This is the first for 4-bit MCU calculator, also is the first personal ability to afford a computer processor!
4004 contains 2300 transistors, functionality is rather limited, but very slow, then blue Giants IBM and most business users to scoff. But it is a landmark product, from then on, INTEL it up with microprocessor. It can be said that the development history of the CPU in a certain sense it is Intel Corporation x86 series CPU.
In 1974, Intel Corporation and on the basis of the 8008 developed 8080 processor, has a 16-bit address bus and 8-bit data bus, contains seven 8-bit register (A, B, C, D, E, F, G, BC, DE, HL combo may consist of 16-bit data register), support 16-bit memory, it also contains some input and output ports, this is a very successful design, but also an effective solution to the external device memory addressability.
In 1978, Intel Corporation has once again lead the way, the first production out of the 16-bit microprocessor, and named the i8086, also produced with the support of the math coprocessor i8087, these two chips use compatible instruction set, but concentrated in i8087 directive added some specialized for logarithmic, exponential and trigonometric functions, math instruction.
In 1979, the Intel Corporation has developed the 8088. 8086 and 8088 in the chip are made of 16-bit data transfers, it is called a 16-bit 8086 microprocessor, but each cycle can transmit or receive a 16-bit data, using only each cycle 8088 8 bits. Because most of the original equipment and chip is an 8-bit, 8-bit 8088 external data transmission, reception and these devices are compatible.
8086 and 8088 shortly after the advent of the Intel Corporation began to improve, they will be more integrated on the chip, which was born 80186 and 80188. These two microprocessors internal work with 16-bit, in the external input output 80186 16, like 80188 and 8088 is 8 bits.
In 1981, the United States IBM 8088 chip for its will be developed for the PC, thus creating a brand new computer era. It is started from the 8088, the personal computer (PC) concept in the world. From 8088 to IBM PC, personal computer really go into people's work and life, it also marks the beginning of a new era.
In 1982, the Intel Corporation in the 8086, developed on the basis of the microprocessor, the microprocessor 80286 maximum clock speed to 20MHz, internal and external data transfer is used for 16-bit, 24-bit memory addressing, memory addressability to 16MB.
8086 ~ 80286 this era is the personal computer era, was started in domestic use or even seen PC people seldom, in people's minds is a mysterious thing. To the early 1990s, the country began to popularize the computer.
In the spring of 1985, Intel Corporation has become a first-rate chip company, its commitment to fully developing the next generation of 32-bit core CPU — 80386. Intel to 80386 designed three techniques: using the "class of 286" microprocessors structure, developing 80387 enhanced floating-point arithmetic ability, cache memory speed bottleneck.
1985, October 17, Intel epoch-making products — 80386DX officially released, its interior contains 27.5 million transistors, the clock frequency to gradually improve after 12.5MHz, to 20MHz, 25MHz, 33MHz, finally a small amount of 40MHz products. Due to the 32-bit microprocessors powerful computing capacity, PC applications extended to many areas such as commercial office and computing, engineering and computing, data center, personal entertainment. 80386 enable 32-bit CPU has become the industry standard PC.
In 1989, our familiar 80486 chip launched by Intel. This after four years¡¯ development and $ 300 million investment in chip's greatness lies in its first real break 100 million transistors, integrated 120 million transistors, use 1 micron manufacturing process. 80486 clock frequency from 25MHz gradually increase to 33MHz, 40MHz, 50MHz.
With the continuous development of chip technology, the frequency of CPU faster and faster, and PC peripherals by technology limitations, can withstand working frequency is limited, and this impeded the CPU clock speed. In this case, the CPU multiplier technology, which allows the CPU operating frequency for the microprocessor FSB of 2 ~ 3 times, 486 DX2, 486 DX4 name is resulting.
In 1993, a new generation of comprehensive beyond 486 and 586 CPU, in order to get rid of 486 era name confusion of microprocessors, Intel Corporation to next-generation product named Pentium (Pentium) to distinguish between AMD and Cyrix products. AMD and Cyrix also launched the K5 and 6x86 microprocessor to deal with the chip giant, but as the best performance of the Pentium microprocessors, Intel has accounted forAccording to most of the market.
In order to increase the PC in multimedia and 3D graphics applications, many of the new instruction set came into being, one of the most famous of the three is Intel's MMX, SSE and AMD's 3D NOW!. MMX (MultiMedia Extensions, multimedia extensions) is Intel in 1996, the invention of a multimedia instruction enhancement technology, including 57 multimedia instructions that can handle more data, MMX technology in collaboration with the software, you can get better performance.
Many Pentium (Pentium MMX) the official name is "Pentium with MMX technology", is published in late 1996. Starting from many Pentium, Intel on its start locks manufactured by CPU frequency, but MMX CPU hyper-FSB ability particularly strong, but also by improving the core voltage to hyper-frequency-doubled, so that's when overclocking is a very fashionable. Overclocking the words from the beginning of the pandemic. 3. CPU manufacturers
Intel company initially taken to be named by Moore, "Intel" this word is "Integrated Electronics (integration/e)," two English words forming a synthesis of, the symbol of the new company's main business is in the IC market. Intel is the production of x 86 architecture CPU's big brother, it occupies 80% of market share, production of Intel CPU became the de facto x86CPU technical norms and standards.
AMD (Advanced Micro Devices <高级微型仪器>) ultra semiconductor company IBM and Cyrix
IBM is the International Business Machines Corporation <公司> (international machine company) abbreviation, which is the short name. IBM server chip has always occupied a dominant position, but for a computer chip still could not open market share. And Cyrix Corporation after the merger, so that it finally has its own x 86 chip production line, its production will increasingly sophisticated and complete. Now the MII performance is also good, especially its price is very low.
IDT Corporation
IDT is the processor manufacturer's rising star, but it's not too ripe.
In fact, idt and cyrix has been China's Taiwan via via company acquisitions. Current via the company's cpu product is primarily intended for embedded devices. Apple computer's CPU manufacturers
By IBM supplier main server CPU manufacturers
Main server CPU's including: IBM, HP, and SUN. CPU manufacturer in China
China has always been quite a number of chip manufacturers, but in general PC chip market has no market share. 0520 is 80 's own production of 286 compatible chip. Now Godson, chip also began to enter the generic PC market. 4. CPU key performance indicators are:
For a CPU, performance is strong is it can survive in the market, the first thing to do, then the CPU performance is determined by what factors? here listed in them affect CPU performance of major technical indicators: clock speed
CPU clock speed, i.e. CPU core of clock frequency (CPU Clock Speed). Commonly referred to as a certain number of MHz CPU is, and this is the "number of MHz CPU's clock speed." Many people think that the CPU clock speed is the speed, not necessarily. CPU clock speed represents the CPU in the speed of the pulse signal shocks, and CPU actual computing capacity and there is no direct relationship. Clock speed and the actual operation speed there is a certain relationship, but is still not a sure formula to quantitative numerical relationship between the two, because the CPU operation speed depends on the CPU's line of performance metrics (cache, instruction sets, the number of bits CPU, etc.). As the clock speed does not directly represent the operation speed, so in certain circumstances, it is likely to occur with high frequency of the CPU the actual operation speed is low. For example, AMD AthlonXP series CPU company can mostly to a lower clock speed, Intel's Pentium 4 series CPU more clock CPU performance, so only for AthlonXP CPU series PR value of a name. So the clock speed is only one CPU performance, and not representative of the overall performance of the CPU.
CPU clock speed is the speed of the CPU, but increasing clock speed to improve CPU operation speed is crucial. For example, suppose a CPU in a single clock cycle execution of an operational directive, when the CPU runs at 100MHz clock speed, the ratio of it running on faster when 50MHz frequency doubled. Because 100 clock cycles than 50MHz clock cycles elapsed time cut in half, which is the work of the CPU core frequency at 100MHz performs an arithmetic instruction time required to work only as 10ns than in time 20ns 50MHz clock speed has been reduced by half, the natural operation speed is fast. But the overall speed of the computer not only depends on CPU speed, and other operation of the subsystem concerned, only the frequencies at the same time, the subsystem operatingspeed and data between the various subsystems transmission speed can be improved, the overall speed of the computer can really be improved.
Work to improve CPU clock speed of the production process of the main subject. Since the CPU is in the semiconductor wafer manufacturing, in the silicon chip components need to be joined, as a conductor in high State requirements for wire thinner as short as possible, so as to reduce stray capacitance and wire interference in order to ensure the correct CPU operations. Therefore the manufacturing process, CPU clock speed development is one of the biggest impediments. Front-end totalLine
Bus is putting it in one or more source parts are delivered to one or more destination part of a group of transmission lines. Pop said, this is more than one part of the public, used to transfer information between the various parts. People often express the speed in MHz to describe the bus frequency. Many types of bus, FSB English name is Front Side Bus, usually with FSB CPU that will connect to the Northbridge bus. The computer's front side bus frequency is determined by the CPU and Northbridge common decisions.
North Bridge chip memory, video card, for contact information such as the throughput of the greatest parts, and southbridge chip connection. CPU is through the front side bus (FSB) to connect to the Northbridge, and thus through the North Bridge chip and memory, video card Exchange data. FSB is the CPU and the outside world to Exchange data at the primary channel, front-side bus data transfer capability on your computer's overall performance effect is very large, if not enough fast front side bus, and powerful CPU can significantly improve the overall speed of your computer. Maximum bandwidth data transmission depends on all at the same time, the width of the data transmitted and the transmission frequency, i.e. data bandwidth = (bus frequency × data bits wide) ÷ 8. Current PC can achieve the FSB frequency have 266MHz, 333MHz,400MHz, 533MHz, 800MHz several larger, FSB frequency, represents the CPU and Northbridge data transmission capacity between bigger, better able to fully exploit the capabilities of the CPU. Current CPU technology development soon, soon, faster operation and the FSB is large enough to ensure adequate data supply to CPU, a lower FSB will not be able to supply enough data to CPU, this limits the CPU performance have to play, to become system bottlenecks. FSB
FSB is the CPU and the whole computer system unit of the base frequency is MHz (megahertz). In the early days of computers, memory and motherboard running speed equal to the FSB, which can be understood as the CPU FSB is connected directly with the memory, for the synchronization between the running state. For the current computer system, they can do the same, but the significance of the FSB still exists in computer systems are most often on the basis of the FSB, multiplied by a certain ratio, the ratio is greater than 1 or less than 1.
Speaking of processor FSB, we must mention the closely related concept: double frequency and clock speed, clock speed is the CPU clock frequency; multiplier that clock speed and FSB of multiples. 1 GHz, FSB, multiplier, its relation: clock speed = FSB x multiplier. Multiplier
CPU multiplier, the full name is the multiplier coefficient. CPU core frequency and FSB a ratio between, this ratio is the multiplier coefficient, hereinafter referred to as multipliers. Theoretically doubling from 1.5 to unlimited, but it is important to note that the double frequency is as an interval unit 0.5. FSB and multiplied by the multiplier is the clock speed, so any improvement you can make the CPU clock speed increase.
The original concept was not harmonic, CPU clock speed and system bus speed is the same, but the CPU's speed faster and faster, the multiplier technology also answered. It makes the system bus to work at relatively low frequencies, and CPU speeds through the multiplier toincrease indefinitely. Then the CPU clock speed formula becomes: clock speed = FSB x multiplier. The double frequency refers to the CPU and system bus between multiple, when FSB unchanged, increase the multiplier, the higher the CPU clock speed also.
FSB and front side bus frequency difference: front side bus speed refers to the CPU and Northbridge-bus speed, more substantive represents the CPU and the outside world data transfer speed. The concept of the FSB is built on the Digital pulse oscillation speed on the basis of, say, 100MHz FSB especially Digital pulse signal on a per-second shock, it was more of a PCI and other bus frequency. The reason why FSB and FSB this confusing two concepts, the main reason is that in the past for a long time (mainly in the Pentium 4 before and emerging Pentium 4), front side bus frequency and the FSB is the same, so often direct says FSB for the FSB, the end result of this misunderstanding. As computer technology development, people found front side bus frequency needs above the FSB, adopted the QDR (Quad Date Rate) technology, or other similar technology to achieve this objective. Principle of these technologies similar to AGP's 2X or 4X, they make the front-side bus frequency a FSB of 2 x, 4 x or even higher, since then FSB and FSB difference began to be attention. Cache
CPU cache (Cache Memory) is located between the CPU and memory temporary memory, it is small but Exchange memory speed. The data in the cache is the memory of a small part of it, but this is a small portion of the CPU is within a short period of time, when the CPU access calls the large amounts of data, you can avoid the memory directly from the cache, thus speeding up the reading speed. Thus, in the CPU's cache is an efficient solution, so that the entire internal memory (cache + memory) becomes both a cache of high speed and memory of large-capacity storage systems. Cache on CPU performance impact is very large, mainly because of CPU data exchange orderand CPU and cache bandwidth.
Most earlier CPU cache is a holistic, but capacity is low, the Intel Corporation from Pentium era started to cache categorized. Was integrated in the CPU core in the cache are no longer adequate for CPU requirements, and restrictions on the manufacturing process and does not significantly improve the capacity of the cache. Therefore there was integrated in the same field with the CPU board or motherboard cache, in which case the CPU core integrated cache called the L1 cache, and external called L2 cache. Level cache is also divided into data cache (Data Cache, D-Cache) and instruction cache (Cache, I-Cache Instruction). Both are used for storing data and instructions for performing these data, but they can also be CPU access, reduces contention conflicts caused by the Cache, improves processor efficiency. Core type
Core (Die), also known as the kernel, is the most important component of CPU. Centre for the uplift CPU chip is the core, from silicon to certain production process manufacturing, CPU all computed, acceptance/storage order, processing data from the core implementation. Various CPU core with fixed logical structure, L1 cache and L2 cache, execution units, instruction-level modules and bus interfaces such as logical units will have the layout of the science.
In order to facilitate the CPU design, production, sales management, the CPU manufacturers of various CPU core gives the corresponding code number, which is the so-called CPU core type.
Different CPU (different series or series) will have different core type (e.g. Pentium 4 Northwood and Willamette and K6-2-K6-CXT and 2 + ST-50, etc.), or even the same core will have different versions of the type (for example, Northwood core is divided into B0 and C1, etc.), the core version of the change is to fix the previous edition of some errors and improve certain performance, but these changes ordinary consumers rarely pay attention to. Generally speaking, the new core types are often better than the old core types have better performance. Therefore, the core types to some extent, determine the CPU work performance. Fabrication process
Usually we say CPU "craft" means a production CPU process, to carry out processing of various circuits and electronic components, manufacturing wire connection of individual components. Typically the accuracy of its production to micron (unit of length, equal to one thousandth of a mm 1 micron) to indicate that future Nano (1 nanometer is equal to one thousandth micrometers) development trends, higher precision, more advanced manufacturing process. In the same materials can create more electronic components and connectors are also thinner, increasing the CPU's integrated, CPU power consumption.
Manufacturing technology of Micron IC refers to the internal and the distance between the circuit. Manufacturing technology trend is, the higher the intensity. Density, the higher the level of IC circuit design, means that the same-sized area of IC, you can have a density higher and more complex circuit design. Microelectronics technology's development and progress, mainly rely on the continuous improvement of the technology that allows devices feature sizes continue to shrink, thereby increasing integration, less power, the device performance is improved. Chip manufacturing process in 1995 from 0.5 Micron, 0.35 Micron, 0.25 Micron, 0.18 Micron, 0.15 Micron, 0.13 Micron, 0.09 Micron, and Micron 0.065 (65nm) manufacturing process will be the next-generation CPU's development goals.

Packaging technology
The so-called "packaging" is an integrated circuit with insulated plastic or ceramic packaging technology. A CPU, for example, we see the appearance of volume and is not a real CPU core size and appearance, but is a CPU core, and other components of the product after the encapsulated.
Package is required for the chip, is also critical. Because the chip must be isolated from the outside world toprevent air impurities on chip circuit of corrosion caused electrical performance. On the other hand, the package after the chip is also easier to install and transport. Due to packaging technology or bad also directly affects the performance of the chip itself play and connected to the PCB (printed circuit board) for the design and manufacturing, therefore it is essential. Package also includes installation of semiconductor integrated circuit chip used in the enclosure, it not only plays placement, fixed, seal and protect the chip and enhanced thermal performance, but also communication chip internal world and external circuits of the bridge-chip contact wire connected to encapsulate shell pin, the PIN through the printed circuit board conductors and establish a connection to other devices. Therefore, for a lot of integrated circuit products, packaging technologies are key. Hyper-Threading technology
Hyper-Threading technology is one of my CPU at the same time perform more procedures and to share a single CPU resources within, in theory, like two CPU at the same time as the execution of two threads, P4 processor needs to add a Logical CPU Pointer (logical processing unit). Therefore a new generation of P4 HT's die area than previous P4 increased 5%. While the remaining part of ALU (integer arithmetic unit), the FPU (floating point unit), the L2 Cache (L2 cache) remain unchanged, the section is being shared.
Although using Hyper-Threading technology can simultaneously execute two threads, but it's not like two real CPU, the CPU has each separate resources. When two threads at the same time requires a certain resource, one that you want to temporarily stop and yield the resources until these resource idle before continuing. So hyper-threading performance does not mean that the two CPU performance.
It should be noted that, with Hyper-Threading technology CPU needs chipset, software support, you can compare the ideal play the technology.

Dual-core typebr>In 2005, clock speed has always been two big processor giants Intel and AMD competing to chase focus. And processor clock speed is also on the Intel and AMD's efforts reached a peak in the processor clock speed lifting speed, also found in the present case, a simple clock speed upgrade has failed to improve the overall performance of the system offers significant benefits, and will bring great heat, the processor disadvantage is the Intel and AMD processor clock speed of the two has some great promotion. In this case, Intel and AMD are not aboutWhile at the same place will be directed to a number of core development direction without conducting a large-scale development of existing products into a theoretical performance more-powerful multi-core processor systems, undoubtedly is a very wise choice.
Dual-core processors on a single semiconductor processor with two different features of the processor core, that is, the two physical processor core integration into a kernel. In fact, the dual-core architecture and not a new technology, however the previous dual-core processor has been Server patent, now began to spread. 64-bit technology
Here's a 64-bit technology is relative to the 32-bit, this bit refers to the CPU GPRs (General-Purpose Registers, general-purpose registers) data width are 64-bit CPU in unit time can handle binary digits. 64-bit instruction set that is running a 64-bit data of the directive, which means that the processor can run at one time 64bit data. Calculation of main 64bit has two major advantages: you can make a wider integer arithmetic; can support greater memory.