【 Weak current College 】 video analysis of common technical terms (1)

【 Overview 】
The video card as computer hosts an important part of the commitment of the output display graphics tasks, like playing games and engaged in professional graphic design is very important. Current civilian video graphics chip supplier in the ATI and nVIDIA mainly consists of the two.

【 Working principle】
Information (data) when leaving the CPU, must pass through four steps, finally reach the display screen:
1. from the bus (bus) into the graphics chip-CPU sent information to the graphics chip inside for processing. (Digital)
2. enter from videochipset videoRAM-will chip processed information sent to the video memory. (Digital)
3. from memory into DigitalAnalogConverter (= RAMDAC), read by display memory out of the information and then sent to a data conversion of RAMDAC (digital to analog). (Digital)
4. from the DAC to monitor (Monitor)-the transformation of analogical data sent to the display (analogical)
Show effectiveness is part of the system's effectiveness, their effectiveness of the above four steps may determine that the effectiveness of the display card (videoperformance) not too, if you want to strictly separate, display card of effectiveness should be determined by the middle two step, because the two steps of data transmission are the internal display card. The first step is the CPU (calculator and controller together make up the core of a computer, a microprocessor or CPU, the CPU) entry to display the card inside, the last step is the display card directly to the screen.

【 Basic structure 】
1）GPU
Full name is GraphicProcessingUnit, Chinese translation for "graphics processor. NVIDIA GeForce256 publishing company in graphic processing chip first initiated the concept. GPU makes video reduces dependence on the CPU, and make part of the original CPU work, especially when in 3D graphics. The GPU core technology with hardware T & l, cubic environment texture and vertex blending, texture compression and a bump map texture, dual texture four pixels, 256-bit render engine and hardware T & l technology can be said to be the sign of the GPU.
2) display card
Display card (DisplayCard) basic function is to control your computer's graphics output, display card to connect a monitor, we can see on the display screen, the display card images display chip, display memory, RAMDAC, etc, which determines the output on your computer screen, including the screen image display speed, color, and display resolution. Display card from the early days of monochrome display card, color display card, reinforced the drawing display card, up to VGA (VideoGraphicArray) displays the drawing arrays, are dominated by IBM display card specifications. VGA in text mode to 720 * 400 resolution, in the drawing mode for a 640 * 480 * 16 colors, or 320 * 200 * 256 colors, 256 colors display mode becomes later display card of common standards, therefore we known as display card to VGA. But then each display chip manufacturers more committed to VGA display capacity and upgrading, and SVGA (SuperVGA), XGA (eXtendedGraphicArray) nouns appear, in recent years shows the chip manufacturers even 3D features and VGA, becomes our current sequential called 3D accelerator card, 3D drawing display card.
3) pixel fill rate
Pixel fill rate of the maximum value is rendered 3D clock multiplied by the number of ways. As NVIDIA's GeForce2GTS chip core frequency for 200MHz, 4 rendering pipeline, each rendering pipeline contains 2 texture units. It's fill rate for 4x2 x2 million pixels/second = 16 million pixels/sec. Composed of pixels here we see on the display screen at 800 x 600 resolution in total have 800x600 = 480,000 pixels, and so on 1024 x 768 monitor resolution with 1024x768 = 786,432 pixels. We play the game and some graphics software often set the resolution, the higher the resolution, the chip will appear when rendered with more pixels, the size of the fill rate for measuring the performance of a video card is important. We calculated the GTS fill rate is 16 million pixels/second, let's see MX200. It's a standard core frequencies rendering pipeline only 175, 2 bars, then it's fill rate 2x2 x1.75 million pixels/second = 7 billion pixels per second, this is it different than the GTS performance half an important reason.
4) memory
Displays a short memory. As the name suggests, its main function is to temporarily store display chip to process data and data processing is complete. Graphics core, the stronger the performance, you need graphics memory. Previous video memory of mainly SDR, capacity is not large. Now available in the Basic is DDR specification, and in some high-end card is used the performance even more outstanding DDRII or DDRIII generation memory (for the moment, DDRIII has not more excellent, but the worst kind).

【 Product 】
1) AGP interface
AccelerateGraphicalPort is Intel Corporation a video interface standard, is designed to address PCI bus low-bandwidth and the development of interface technology. It will be the graphics card and connect the system's main memory, CPU and graphics processor directly between opened up a faster bus. Its development has undergone AGP1.0 (AGP1X/2X), AGP2.0 (AGP4X), AGP3.0(AGP8X)。 The latest theoretical bandwidth for AGP8X its 2.1Gbit/seconds.
2) PCIExpress connector
PCIExpress is the next generation of bus interface, and the use of such interface graphics products, has formally launched in 2004. Early in the spring of 2001, "Intel Developer Forum, Intel Corporation raises to replace with a new generation of technology and a variety of chip PCI internal connections and call it a third-generation I/O bus technology. Then at the end of 2001, including Intel, AMD, DELL, IBM, over 20 industry-leading companies to begin drafting a new technical specifications and completed in 2002, on its official named PCIExpress. Theoretical speed of up to 10Gbit above, so the gap, the AGP has been hit by the almost PCIE, but like PCI instead of ISA, it requires some time, and must be a 915 more North Bridge support PCIE, so that you can anticipate the need to replace the AGP also PCIE for a long time.
3) now the hottest double card technology
SLIScanLineInterlace (scan-line Interleave) technology is applied to the Voodoo 3dfx company's technology, it passed the 2 block Voodoo card physical connection with SLI cable, work a Voodoo card is responsible for rendering screen odd line scan, another is responsible for rendering even row scan, so as to achieve two video "get connected" with the "double" of performance.

【 Software configuration 】
1）DirectX
DirectX is not a simple graphics API, which is Microsoft's development of a wide range of uses of the API, it contains DirectGraphics (Direct3D + DirectDraw), DirectInput, DirectPlay, DirectShow, DirectSound, DirectSetup, DirectMediaObjects and other components, it provides a complete set of multimedia interface programmes. Is it in 3D graphics, excellent performance, it is looked for other areas. DirectX development beginning in order to compensate for Windows3.1 system on graphics, sound processing, and have now become on the whole the various aspects of the multimedia system has a decisive influence on the interface.
2）Direct3D
To tell Direct3D cannot but speak the DirectX, DirectX is a Microsoft development and publishing of multimedia development package, part of which is called the DirectDraw API is a graphic drawn inversion, provides a powerful graphics processing capability of access, and integrate the DirectDraw some 3D graphics-related features, called Direct3D. Probably because it is Microsoft's hand, some people say it will become the standard for 3D graphics.
3）OpenGL
OpenGL is an abbreviation, OpenGraphicsLib is a 3-d graphics library, is also the industry standard in this area. Computer 3D graphics is used data description into three-dimensional space through calculation of two-dimensional images and display or print technologies. OpenGL is supported this conversion library, it comes from SGI for their graphics workstation development IRISGL, cross-platform porting process developed into OpenGL. SGI in July 1992 after publishing 1.0, become the industry standard, which was founded in 1992 by the independent Consortium OpenGLArchitectureReviewBoard (ARB) control. SGI, ARB members voted to produce standard and made into a specification document (Specification) announced that the software and hardware manufacturers according to this developing your own system. Only through a full test of ARB specification for OpenGL can be called. In December 1995, approved the 1.1 ARB, latest version specification version is 1.2.1. adopted 1999.5

【Main parameter 】
1. display chip (model number, revision levels, develop code, manufacturing process, the core frequency)
2. the memory (types, bandwidth, capacity, package types, speed, frequency)
3. technology (pixel rendering pipeline, the vertex shader engine number, 3DAPI, RAMDAC frequency and support MAX resolution)
4.PCB Board (PCB layers, graphics interface, output interface, the heatsink)

1) display chip
Display chip:
Also known as the figure type processor-GPU, it's role in the video, as CPU in the role of the computer. More direct analogy is the brain in the role of the human body.
First briefly common production display chip vendor: Intel, ATI, nVidia, VIA (S3), SIS, Matrox, 3DLabs.
Intel, VIA (S3), SIS produces integrated chip;
ATI and nVidia to independent chip, is currently on the market, but due to AMD ATi now has been acquired, whether will continue after the independent display chip is hard to say;
Matrox, 3DLabs is primarily intended for professional graphics market.
Because of the ATI and nVidia video card basic occupy the mainstream market, the following main will focus on the two companies ' products.
Model:
ATi Radeon company's primary brands (Radium) series, its model from its RadeonXpress200 to Radeon (X300, X550, X600, X700, X800, X850) to recent
> Radeon (X1300, X1600, X1800, X1900, X1950) performance followed by low to high.
NVIDIA Corporation's main brand GeForce series, its model from the GeForce256, GeForce2 (100/200/400), GeForce3 (200/500), GeForce4
(420/440/460/4200/4000/4400/4600/4800) GeForceFX (5200/5500/5600/5700/5800/5900/5950), GeForce
(6150/6100/6200/6500/6400/6600/6800/), GeForce (7300/7600/7800/7900/7950) to recent GeForce (8400/8500/8600/8700/8800) performance followed by low to high.
Version level:
In addition to the standard version, but also some special edition, Special Edition will normally be in the standard version of the model followed by a suffix, common are:
ATi:
SE (SimplifyEdition Lite) is usually only 64bit memory interface, or a reduction in the number of pixel line.
Pro (ProfessionalEdition Professional Edition)-frequency version, General version than the standard number in the line number and frequency/points those aspects have to be slightly higher.
XT (eXTreme high-end Edition) is a series of high-end ATi, and nVIDIA for low-end models.
XTPE (eXTremePremiumEditionXT Platinum Edition) high-end models.
XL (eXtremeLimited high-end series of lower-end models) ATI's newest version of high frequency R430
XTX (XTeXtreme high-end version) published after X1000 series new naming rules.
CE (CrossfireEdition CrossFire Edition) crossfire.
VIVO (VIDEOINandVIDEOOUT) means a graphics card with video input and video capture two functions.
HM (HyperMemory) could take up memory video card
nVIDIA:
Based on the XT ZT again block to lower prices.
XT block version, while in most high-end ATi said.
LE (LowerEdition low-end version) and XT Basic, ATi is also used.
Version MX parity, public class.
GTS/GS-frequency version.
GE is a strong point, GS is hyper-the frequency of GS. /Yingchi video card is used to indicate the ashes player Edition "stuff
GT-frequency version. More than one grade GS high because there is no shrinking pipeline and GT vertex element.
GTO is slightly stronger than the GT, a bit of a vehicle within the point of the GTO.
Ultra in GF7 series represents the most high-end before, but the 7 series top-end GTX named it replaced.
GTX (GTeXtreme) Enhanced Edition, block or reduce water pipes as GT, and then continue to shrink as GS version.
GT2 Dual GPU graphics card.
TI (Titanium and titanium) General is the representative of the nVidia's high-end versions.
Go for mobile platforms.
TC (TurboCache) can take up memory video card