显示适配器（Display adapter）

显示适配器（Display adapter） 显示适配器(Display adapter) Graphics (Video card, Graphics card), also known as the display adapter (Video adapter), is the most basic part of personal computer. The purpose is to convert the graphics display driver information display to the computer system, and provides a scan signal to the display and display correctly display control, is an important element connected with the display and the personal computer motherboard, "is one of the important equipment in human-computer dialogue". The card is an important part of the host computer, undertake the task to output and display picture, play games and love for people engaged in professional graphic design graphics is very important. The main composition of graphics (and parameters) 1, display chip (model version, level, development code, manufacturing process, core frequency) 2, memory (type, width, capacity, speed, frequency, package type) 3, technology (pixel shaders, vertex shader, 3D API, RAMDAC frequency and support MAX resolution) 4, PCB board (PCB layer, graphics interface and output interface, radiating device) 5, brand 1, display chip Display chip, also called pattern processor - GPU, its role in graphics, just as the role of CPU in the computer. The direct effect of metaphor is the brain in the human body. Take a brief look at the common production display chip manufacturers: Intel, ATI (acquired by AMD, the brand changed to AMD), nVidia, VIA (S3), SIS, Matrox, 3D, Labs. Intel, VIA (S3), the main production of integrated chip SIS; ATI (AMD), nVidia with independent chip, is currently on the market mainstream, but because the ATi has been acquired by AMD, still continue to develop independent display chip; Matrox, 3D Labs is mainly for professional graphics market. Because ATI and nVidia occupy the mainstream graphics market, the following will be the main focus of the two companies are introduced. Model The company's main brand ATi Radeon (RA) series, the type of the Radeon Xpress from early 200 to Radeon (X300, X550, X600, X700, X800, X850) to Radeon (X1300, X1600, X1800, X1900, X1950), in the latest HD series (HD3650/3850/4850/4870 etc.). The main brand GeForce series of nVIDIA company, the model by the GeForce as early as 256, GeForce2 (100/200/400), GeForce3 (200/500), GeForce4 (420/440/460/4000/4200/4400/4600/4800) to GeForceFX (5200/5500/5600/5700/5800/5900/5950), GeForce (6100/ 6150/6200/6400/6500/6600/6800/), GeForce (7300/7600/7800/7900/7950), GeForce 8 series, GeForce 9 series of the latest GTS/GTX 2XX series of high-end. Release level In addition to the Standard Version, some special edition, special edition will generally be in the standard version of the model followed by a suffix, are common: ATi: SE (Simplify Edition simplified version) is usually only 64bit memory interface, or reduce the number of pixel lines. Pro (Professional Edition Professional Edition) high frequency version, generally higher than the standard version in the line number / the number of vertices and the frequency of these are a little higher. XT (eXTreme high-end version) is a high-end ATi series, and nVIDIA used as the low-end models. XT PE (eXTreme Premium Edition XT Platinum Edition high-end models). XL (a eXtreme Limited series of high-end low-end models in high frequency) launched the latest version of ATI in R430 XTX (XT eXtreme X1000 series of high-end version) after the release of the new naming rules. CE (Crossfire Edition crossfire crossfire edition). VIVO (VIDEO IN and VIDEO OUT) refers to the card with video input and video capture two functions. HM (Hyper Memory) can use memory cards NVIDIA: ZT based on the XT frequency to reduce prices fall again. XT frequency version, while ATi said the most high-end. LE (Lower Edition and XT as the basic low-end version), also used ATi. MX cheap version, public class. GTS/GS low version. GE better than GS, is actually the ultra frequency of GS. GT frequency version. A grade higher than GS because GT did not reduce the pipeline and vertex unit. GTO better than GT, a car in the taste of GTO. Ultra represents the most high-end GF7 series before, but the 7 series of the most high-end name changed to GTX. GTX (GT eXtreme) enhanced version, frequency reduction or reduction of pipeline into GT, and then continue to shrink into GS version. GT2 dual GPU graphics. TI (Titanium Ti) is generally representative of the high-end version of nVidia. Go multi language mobile platform. TC (Turbo Cache) can use memory cards Codename The codename is display chip manufacturers in order to facilitate the unified display chip in the design, production, sales and management driven architecture of display chip corresponding basic code of a series of. Codename role is to reduce the display chip manufacturer cost, rich product line and achieve a unified driver. In general, display chip manufacturers can use a basic development code through quantity control, pipeline number, vertex shader unit memory type, memory interface, core and memory frequencies, supported by technical characteristics so as to derive a series of display chip to meet different performance, price, market etc. that can also have some flaws in the process of manufacturing high-end display chip products become the corresponding low-end fully qualified display chip products sold by shielding pipeline and other methods, can greatly reduce the difficulty and cost of design and manufacturing, enrich its product line. With the development of a code display chip can use the same driver, the display driver chip manufacturers and consumers to use cards are provided for convenience. Rendering architecture display chip with a development code and technical characteristics of support is basically the same, but the process is the same, so the development of code is an important parameter in evaluating the performance of graphics and grades. Different versions of the same type can be a code, for example: GeForce (X700, X700 Pro, X700 XT) code are RV410; Radeon (X1900, X1900XT, X1900XTX) code is R580, but there are other things, such as: GeForce (7300 LE, 7300GS) code is G72 GeForce; and (7300 GT, 7600 GS, 7600 GT) code is G73. manufacturing process Refers to the manufacturing process in the production process of GPU, for processing a variety of circuit and electronic components manufacturing, wire connecting various components. Usually the production of precision to UM (micron) to represent the future to the development trend of nm (nm) (1mm=1000um 1um=1000nm), higher accuracy, more advanced production technology. We can manufacture more electronic components in the same material, the connection line is fine, improve the integration of the chip, the chip power consumption is small. Micro manufacturing technology is refers to between the circuit and IC circuit in the distance. Manufacturing is the trend of development of high concentration to the direction of. IC circuit design of higher density, means that in the same area the size of the IC, can have higher density and more complex functions of the circuit design. The development and progress of the microelectronics technology, mainly rely on technology improvement, the device feature size continues to shrink, thus increasing integration, lower power consumption, improve device performance. The chip manufacturing process in 1995, from the 0.5 micron, 0.35 micron, 0.25 micron, 0.18 micron, 0.15 micron, 0.13 micron, 90 nm to the current mainstream (0. 09 nm), 65 nm, 45 nm. Core frequency The core frequency refers to the frequency display graphics core, its working frequency can reflect the performance of the core to a certain extent, but the graphics performance is decided by the core frequency, memory, pixel lines, pixel fill rate and so on many aspects, so the core of the show under different core frequency this card does not mean strong performance. For example, the core frequency of 9600PRO reached 400MHz, 9800PRO than 380MHz, but the performance of absolute 9800PRO is stronger than 9600PRO. The same level of the chip, the core of the high frequency performance is stronger, the core frequency is one of the methods to improve the overclocking. The mainstream display chip is only ATI and NVIDIA two, both provide a display core to the third party vendors in the same core of the show, some manufacturers will display their products to improve the core frequency appropriately, make it work in the above shows the core fixed frequency to achieve higher performance. 2, memory Type: type of memory in the market at present are mainly SDRAM, DDR SDRAM, DDR SGRAM three. SDRAM particles are mainly used in the low-end graphics card, the frequency is generally not more than 200MHz, in terms of price and performance than DDR have what advantage, thus gradually replaced by DDR. DDR SDRAM Double Data Rate is the abbreviation of SDRAM (double data rate), it can provide a high frequency, excellent processing performance data. DDRSGRAM is a graphics card manufacturers especially for drawing requirements, in order to strengthen the graphics processing and access control from the drawing efficiency, synchronous dynamic random access memory (SDRAM) and the improvement of the product. SGRAM allows to block (Blocks) for individual modifications or access the data in memory, it can with the central processor (CPU) synchronization, can reduce the memory access times, increase the efficiency of drawing controller, despite its good stability, and the performance is also very good, but it is poor overclocking performance. At present, the mainstream of the market is DDR2 and DDR3. Width Is the memory interface to transmit data to memory bits in one clock cycle, the greater the amount of data bits can be transmitted more moment, this is one of the important parameters of memory. The width of memory on the market there are 64 bit, 128 bit, 256 bit and 512 bit types, 64 bit graphics, traditionally called 128 bit graphics and 256 bit graphics refers to the corresponding memory interface. Memory interface is higher, the better the performance price is higher, so the 512 bit memory is more suited for the high-end graphics, and mainstream graphics using basic 128 and 256 bit memory. Memory bandwidth = memory frequency X /8 memory interface, the memory frequency quite, memory interface will determine the size of the memory bandwidth. For example: the same memory frequency for 128 bit and 256 bit 500MHz memory, then Talia memory bandwidth respectively: 128 = 500MHz*128 / 8=8GB/s, 256 = 500MHz*256 / 8=16GB/s, is 2 times 128, showing the importance of data in video memory interface. The memory card is composed of a block of memory chip, memory interface is also composed of the total width of memory particles. Memory interface = memory bit * memory particle particle number. The number of memory memory particles are provided with related manufacturers, can go online to find the number, will be able to understand its width, multiplied by the number of memory particles, can get the bus width. capacity This is better understood, the bigger the more the better things, of course. The current mainstream memory capacity, 64MB, 128MB, 256MB, 512MB etc.. Package type The main form of a package: TSOP (Thin Small Package OutLine TSOP) QFP (Quad Flat Package) quad flat package 微型(微型球栅阵列)微型球闸阵列封装，又称FBGA(细间距球栅阵列) 目前的主流显卡基本上是用TSOP封装和封装，其中又以TSOP封装居多。 速度 显存速度一般以NS(纳秒)为单位。常见的显存速度有7ns、6ns、5.5ns、5ns、4ns，3.6ns、2.8ns、2.2ns、1.1ns等，越小示速度越快\越好。 显存的理论工作频率计算公式是:额定工作频率(MHz),1000 /显存速度×N(N得到因显存类型不同而不同，如果是SDRAM显存，则n = 1;DDR显存则n = 2;DDRII显存则n = 4)。 频率 显存频率一定程度上反应着该显存的速度，以兆赫(兆赫兹)为单位。 显存频率随着显存的类型、性能的不同而不同: SDRAM显存一般都工作在较低的频率上，一般就是133mhz和166MHz，此种频率早已无法满足现在显卡的需求。 显存则能提供较高的显存频率DDRSDRAM，因此是目前采用最为广泛的显存类型，目前无论中、低端显卡，还是高端显卡大部分都采用DDRSDRAM，其所能提供的显存频率也差异很大，主要有400mhz、500mhz、600mhz、650mhz等，高端产品中还有800mhz或900MHz，乃至更高。 显存频率与显存时钟周期是相关的，二者成倒数关系，也就是显存频率,1 /显存时钟周期。如果是SDRAM显存，其时钟周期为6ns，那么它的显存频率就为1 / 6ns = 166兆赫;而对于DDRSDRAM，其时钟周期为6ns，那么它的显存频率就为1 / 6ns = 166MHz，但要了解的是这是DDRSDRAM的实际频率，而不是我们平时所说的DDR显存频率。因为DDR SDRAM在时钟上升期和下降期都进行数据传输，其一个周期传输两次数据，相当于频率的二倍。习惯上称呼的DDR频率是其等效频率，是在其实际工作频率上乘以2，就得到了等效频率。因此6ns的DDR显存，其显存频率为1 / 6ns * 2 = 333mhz。但要明白的是显卡制造时，厂商设定了显存实际工作频率，而实际工作频率不一定等于显存最大频率。此类情况现在较为常见，如显存最大能工作在650MHz，而制造时显卡工作频率被设定为550兆赫，此时显存就存在一定的超频空间这也就是目前厂商惯用的，显卡以超频为卖点。 3、技术 象素渲染管线 渲染管线也称为渲染流水线，是显示芯片内部处理图形信号相互独立的的并行处理单元。 在某种程度上可以把渲染管线比喻为工厂里面常见的各种生产流水线，工厂里的生产流水线是为了提高产品的生产能力和效率，而渲染管线则是提高显卡的工作能力和效率。渲染管线的数量一般是以像素 渲染流水线的数量×每管线的纹理单元数量来表示。例如， geforce6800ultra的渲染管线是16×1，就表示其具有16条像素渲 染流水线，每管线具有1个纹理单元;geforce4mx440的渲染管线是 2×2，就表示其具有2条像素渲染流水线，每管线具有2个纹理单元 等等，其余表示方式以此类推渲染管线的数量是决定显示芯片性能和 档次的最重要的参数之一，在相同的显卡核心频率下，更多的渲染管 线也就意味着更大的像素填充率和纹理填充率，从显卡的渲染管线数 量上可以大致判断出显卡的性能高低档次。 But the number of graphics performance not only depends on the rendering pipeline, at the same time also depends on the core architecture, the rendering pipeline execution efficiency, and the number of vertex shader unit graphics core frequency and memory frequency etc.. Generally speaking in the same display core architecture, rendering pipeline more also means higher performance, such as 16 x 1 Architecture GeForce6800GT and its performance is better than 12 * 1 GeForce6800 architecture, using the same technology as factory 2 production lines of production capacity and efficiency than the 1 production lines.; in different display core architecture, the number of the rendering pipeline more does not mean better performance, such as 4 x 2 architecture of GeForce2GTS and its performance is inferior to the 2 x 2 architecture GeForce4MX440, like the factory with 1 lines of advanced technology production capacity and efficiency but also stronger than the the old 2 production lines. The number of vertex shader Vertex shader (Vertex Shader), also known as the eclipser vertex, according to the official specifications, vertex shader is a processing unit to increase all kinds of effects in the 3D field in the shadow, vertex shader The programmable feature allows developers to load new software by instructions to adjust all kinds of effects, each vertex will be all the data variables are clearly defined, including X, y, Z at least the coordinates of each vertex, each vertex may envelope data color, the initial path, the material, characteristics of light. The number of vertex shader faster. 3D API API is the abbreviation of Interface Application Programming, application program interface means, while the 3D API refers to the interface card and the application directly. 3DAPI allows programmers design 3D software as long as the API calls within the program, so that the API automatic and hardware driver communication, 3D graphics start 3D chip powerful, thus greatly improve the design efficiency of the 3D program. If there is no 3DAPI in the development of the program, the programmer must understand the characteristics of all graphics, can produce completely matched with the graphics program, play all the graphics performance. Is this 3DAPI interface card and software directly, only require programmers to write the line interface code, you can give full play to the graphics card does not need to know the hardware specific performance and parameters, which greatly simplifies the efficiency of program development. Similarly, display chip manufacturers to design their own hardware products according to the standard, in order to achieve the optimization in the API call a hardware resource, better performance. With 3DAPI, we can realize the different manufacturers of hardware and software is compatible with the maximum range. For example, embodied in the most 3DAPI games, game design, do not have to consider the specific characteristics of a card, but only in accordance with the standard 3D API interface to the development of the game, when the game is run directly by calling 3DAPI graphics hardware resources. The application of personal computer in the main 3D API: DirectX and OpenGL. The frequency of RAMDAC and maximum resolution support RAMDAC Random Access Memory Digital/Analog Convertor abbreviation, namely random access memory digital to analog converter. RAMDAC is the analog signal to digital signal conversion memory for the display can show out, the conversion rate of MHz. The computer data processing process is actually the process of digital things, all things will be processed into 0 and 1 number two, and then continuously accumulated. Graphics cards also depend on these 0 and 1 for each pixel of the color, depth, brightness and other processing. The graphics generated a signal is expressed in numbers, but all CRT displays are working in analog mode, the digital signal can be identified, it must have the appropriate equipment will convert digital signals to analog signals. While the conversion of RAMDAC is the graphics in digital signal to analog signal equipment. The conversion rate of RAMDAC to MHz, which determines the refresh frequency (similar to display "bandwidth"). Its working speed is higher, the frequency band is wide, high resolution picture quality is better. When this value determines in enough memory, graphics support the highest resolution and refresh rate. If you want to reach a resolution of 85Hz in the 1024 * 768 resolution, the rate of RAMDAC is at least 1024 x 768 x 85 x 1.344 (conversion coefficient) 83019 106 = 90MHz. The current mainstream graphics RAMDAC can reach 350MHz and 400MHz, have been sufficient to meet and exceed the current most displays can provide the resolution and refresh rate. 4, PCB board PCB is the abbreviation of Block Printed Circuit, also known as printed circuit board. Is the card body (green card board), all the components are on board the PCB, so the PCB board is good or bad, directly determines the quality and stability of electrical performance graphics. The number of The PCB board is generally adopts 4 layers, 6 layers, 8 layers or multiple layers, theoretically less than good, but the premise is based on reasonable design. Each layer of PCB can be divided into the signal layer (Signal), a power supply layer (Power) or ground (Ground). Each layer of the PCB version of the circuit is independent of each other. In the 4 layer of PCB board, the signal layer generally distributed in the PCB of the top layer and the bottom layer, the middle layer and the two is the power and ground plane. The 6 layer PCB is relatively complex, the signal layer generally distributed in 1, 3, 5, and the power supply layer has 2 layers. As for judging the merits of PCB, mainly to observe the printed circuit part is clear, PCB is smooth without any deformation etc.. Graphics interface Common PCI, AGP 2X/4X/8X (now obsolete), is the latest PCI-Express X16 interface, is the current mainstream. Output interface Now the most common primary output interface: VGA (Video Graphics Array) video graphics array interface is to convert the analog signal output good to CRT or LCD display DVI (Digital Visual Interface) digital video interface, a video signal without conversion, no signal attenuation or distortion, the future replacement of the VGA interface. S-Video (Separate Video) S terminal, also called two component video interface, using five general line connector, which is used to output the luminance and chrominance separation equipment, the main function is to overcome the interference of video composite output when the brightness and chroma. Heat radiating device The operation stability of cooling device can also affect the quality of the graphics, there is a common cooling device: Passive cooling: installation of radiator Aluminum Alloy or copper metal etc.. Air cooling: fan installed in the heat sink, most by this method. Water cooling by heat pipe liquid GPU and water pump is generally used in high-end graphics in the top. colour Many people think that red card than the green, green than yellow, and its color and graphics is not what relation, some manufacturers like to use red, some like to use green, this is completely determined by the manufacturer. Some brand-name manufacturers, it is already formed a certain style. Therefore, there is much change in the PCB color generally does not. 5, brand The graphics industry is becoming increasingly fierce competition. All kinds of brand is I think a multitude of names, following some good brands, please don't be too superstitious for reference only: Sapphire (SAPPHIRE), ASUS (ASUS), Hon Hai (Foxconn), Powercolor / Dataland (PowerColor/Dataland), Leadtek (Leadtek), XFX (XFX) and innovision (Inno3D)...... MSI (MSI), Elsa (ELSA), FORSA (FORSA), Germany (Palit), Jie Bo (Jetway), (Abit), abit pzon (EPOX), BIOSTAR (Biostar), gainward (Gainward) and Jing Yu (SPARKLE), GALAXY (GALAXY), Tianyang (GRANDMARS), super Zhuo Tiancai (SuperGrece), (MAXSUN), MAXSUN Xiang Sheng (ASL), yitg (YESTON)......