Video card

A graphics card controls in a computer graphics output. During program execution, the processor calculates the data, it forwards it to the graphics card and the graphics card converts the data so that the monitor or projector can play everything as image. Graphics cards are either a PC expansion cards (via the bus PCI, AGP or PCI Express, formerly ISA or VLB) connected to the motherboard or are included in the chipset on the motherboard. Meanwhile, the integration has progressed, that the component for the graphics output are already in main processors available ( at Intel since Core iX generation of AMD Fusion program).

Other modern high performance video cards are GPU video RAM RAMDAC and the connections for external devices ( eg for the monitor ).

• 3.1 onboard solutions
• 3.2 Business Solutions
• 3.3 Games Graphics Cards
• 3.4 Professional Solutions
• 3.5 Other

• 4.1 software problems with graphics cards

History

The graphics cards principle was first used in series production at the microcomputer Apple II, whose integrated on the motherboard graphics capabilities could be improved by addition to acquiring cards. ( " PAL color card", " 80-column card").

The first IBM PC came in 1981 with a card on the market, which only allowed the monochrome appearance of text (MDA = Monochrome Display Adapter ). The Hercules 1982 offered a much better card, the Hercules Graphics Card.

Until 1989, the color graphics card types translated by default, the new IBM brought to the market:

Even today, the VGA mode (640 × 480 pixels in 16 colors ) of the "emergency mode " on all PCs; only to this mode, the hardware of all modern PC graphics cards can be addressed by the software in a uniform way.

However, VGA was not the last screen size standard. The Video Electronics Standards Association ( VESA) set a standard for video modes up to a resolution of 1280 × 1024 pixels in 16 -bit color depth, which dominates any PC graphics card today.

The other names SVGA, XGA, etc. are no more cards to standards, but short names for screen resolutions, for example, XGA 1024 × 768 points.

Until about 1990, the graphics cards were limited to convert the content of the video RAM via a so-called RAMDAC chip into output signals for the monitor. The programmer was able to use only the text mode and set individual pixel to a specific color in graphics mode essentially. That was the first generation of graphics cards. It was followed by two more:

• The so-called "Windows Accelerator "
• The 3D accelerator

As of 1991, the graphics cards developed into independent small computers with their own GPU ( Graphics Processing Unit), a so-called Graphics or pixel engine or German GPU, where you could not only set individual pixels, but which one commands for drawing lines and could send filling surfaces ( Windows accelerator). These functions accelerated mainly moving the window ( windows ) of the graphical user interface, hence the name. The concept of additional functionality was always continued with time, for example, since 1995, functions for the acceleration of video playback (eg in AVI format ) and decoding of compressed video data (eg, MPEG) were introduced ( video acceleration ). These functions were previously offered on separate cards.

After the mid-1990s had begun with Doom the big boom of 3D games, but soon came out of the first useful 3dfx 3D accelerator, the Voodoo Graphics chipset. A 3D accelerator, a program can specify the geometric figures in the form of polygons and textures in a three dimensional space, with which the surfaces of the polygons will be filled ( rendering). This rather simple but computationally intensive task had yet to take over the CPU in the early 3D games; She could now be delegated to the graphics card, which led to a massive increase in performance of 3D games ( better picture resolution, much more realistic images ).

Were installed on its own plug-in cards by the graphics signal of the built- in system 2D graphics card is looped through the 3D accelerator of the first generation, soon Solutions 2D and 3D functionality prevailed, on a map united.

In order to offer even greater 3D performance, two or more 3D graphics cards or processors connected in parallel in order to calculate more graphic elements per unit time can today with the multi-GPU technology ( see also SLI and AMD CrossFireX ). However, since this technique brings quite high costs and an increased energy demand with herself, she has so far only little influence on the graphics card market.

Design and operation

Hardware interface to the system

The best-known hardware interface for graphics cards are PCI, AGP, and PCI Express, earlier were also ISA or VESA Local Bus consistently. These interfaces are either buses or direct (AGP, PCI Express), which connect the bus controller with the graphics card. Since the specification of the interfaces is usually made ​​by interest groups, in which both the controller and the graphics card or graphics chip maker member, function ( ideally) all compliant graphics card with all compliant controllers. There was in the past, but various problems with some interfaces that restricted interoperability, for example, " AGP Fast Writes" at AGP ( on Intel platforms could increase the power, ensure on AMD platforms for instability) or IRQ problems with PCI (possible crashes / freezing or dips in power, cause mostly poor / faulty implementation of the interface ).

For platforms other than IBM - compatible computers, there were according to the usual there connector systems graphics card for the Apple II slots and later with the first Macs for NuBus (later PCI and assignee as PC), for Amigas for the Zorro bus and also maps of Europe for systems that are based on the latter.

Graphics memory

Graphics memory is used to store the graphics processing unit (GPU ) processed data as well as for video memory ( " framebuffer " ): These are digital images that will later be output on the computer screen or the projector.

The amount of video memory determined the maximum color depth and screen resolution. Here, the memory required for a desired resolution and color depth could be easily calculated by the user: For example, if the resolution is 1600 * 1200 is desired with a color depth of 24 bit (true color), one first calculates the number of dots (pixels ) of this resolution ( 1600 * 1200 horizontal vertical = 1,920,000 pixels in total). The color depth " 24-bit " means that exist for each of these pixels 24-bit color information. Thus, multiplying the number of pixels with the color depth ( 1,920,000 * 24 bits = 46.08 million bits). Now, the conversion in bytes is only required. Because one byte consists of 8 bits, the number is divided by 8 ( 46,080,000 bits / 8 = 5,760,000 bytes). Since earlier graphics cards are usually shipped with 4 or 8 megabytes of video memory, you would have for the desired setting requires a graphics card with at least 8 megabytes of video memory.

Today only graphics cards are built with much more memory than would be required for pure image storage. When rendering three-dimensional graphics are here in addition to the frame buffer, the data of the objects, such as size, shape and position, as well as the textures that are placed on the surface of the objects stored. In particular, the higher and higher resolution textures have caused a sharp increase in memory size in current graphics cards. Thus, the memory size of current graphics cards is already in the higher three - to four -digit megabyte range (256 MB, 512 MB, 1024 MB, 2048 MB, 3072 MB ), 128 MB and less have become rare. In game graphics card is the upper limit in early 2009 at 2 GB, whereas professional graphics cards can be fitted even with up to 16 GB of graphics memory.

With onboard solutions usually the main memory of the system is used as graphics memory, which is referred to as shared memory. Access is via the relevant bus system and is therefore slower than directly tailed memory.

GPU

The graphics processor is used to calculate the screen output. Mid-1990s, the first 3D accelerators came on the market. This graphic processors were capable of some effects and triangle -based algorithms ( such as, inter alia, Z- buffering, texture mapping ), and anti -aliasing to perform independently. Especially the field of computer games helped those in addition to install plug-in cards (such as 3Dfx Voodoo Graphics ) to a surge in development.

Today, GPUs are superior because of their specialization in graphics calculations the CPUs of their computing power. Were used for comparison, the transistor count of the graphics processor from Nvidia ( Geforce 8800GTS 512, 754 million), with a model from Intel ( Core 2 Extreme QX9650, 820 million ). The difference is clear when you consider that consumed more than half the CPU chip area for the 2x6 MB cache. The development of the integration density of GPUs has exceeded an annual factor of 2.4 Moore's Law.

As with the main processors of the computers in the graphics cards are installed, the GPUs are on the graphics cards often the subject of overclocking modifications to improve performance.

The computing capacity available on such graphics cards (see below GPGPU ) has already led to alone to achieve maximum computing power, more graphics cards are installed in a computer. Such systems are partially assembled in large numbers to supercomputers.

Cooling Solutions

Due to the high thermal dissipation caused by the increasing complexity of graphics processors and partly also of the graphics memory similarly expensive cooling solutions like processor coolers are necessary. Graphics cards consume a GPU ( GeForce GTX 680) up to 195 watts ( TDP) power that must be completely dissipated as heat energy. For this purpose, there are several approaches:

• Active air cooling - the thermal energy is dissipated via a heat sink to the ambient air, which is circulated by a fan. This is the simplest and cheapest variant to dissipate large amounts of heat, but also causes noise.
• Passive air cooling - by a heat sink, the thermal energy is dissipated by convection to the ambient air. This is only possible with low power or with very large heatsinks. Often large heat sinks are mounted on both sides of the graphics card, which are connected with a heatpipe. On the problem can result here is that the heat sink can reach a high weight, which leads to a high mechanical load on the socket.
• Water cooling - when water-cooling is used for the CPU, the graphics card can be integrated into this circuit. The thermal energy is then transferred to the water in the circuit and from there through a radiator to the ambient air. This allows transport of large amounts of heat, but is also the most complex and expensive cooling solution. There are also some cards with pre-installed water cooling with the advantage that the warranty remains in effect.

In particular, the constructions of the air cooling systems are required by the surface of the heatsink often much greater than permitted by the specifications of the slot (see figure right ). For this reason, can be used on the mainboard often the adjacent slots not.

RAMDAC

The RAMDAC (random access memory digital / analog converter) is a chip, which is responsible for converting digital ( video memory ) to analog image signals ( monitor). From him the signal outputs are activated. It can also be integrated in the GPU.

External signal outputs

Additional signal outputs and inputs are implemented in different ways depending on the card. Partial appropriate jacks (RCA, S-Video, LFH60 ) are available directly on the bracket. Mainly for reasons of space see manufacturer but also an indirect connection via an adapter cable or breakout cables. Then place directly on the graphics card is a jack, for example, from the mini -DIN family, whose circuitry is not standardized and often has the general name VIVO ( Video -In Video - Out). Here is a vendor-specific breakout cable is connected, which then provides more connection options are available.

Designs and applications

Basically, four types of graphics solutions can be distinguished:

Onboard solutions

These Integrated Graphics Processor, in short IGP solutions mentioned the functionality of the graphics card into the motherboard chipset or the processor (eg, Intel i5 ) integrated. IGPs have all 2D functions, but mostly just slow or outdated 3D functionality and are therefore used primarily in areas with lower graphics requirements. Meanwhile, however, the graphics solutions from AMD and Intel are already powerful enough to play the latest games on at least low details. A lot has been done also in the supported features. Currently support both manufacturers DirectX version 11 and newer technologies such as OpenCL or OpenGL. Due to their low power consumption, they are also often used in notebook computers. The low power consumption is also a motive for use in embedded PCs; for particularly critical applications, such as in medicine is added the advantage that the failure source of the plug contacts between the motherboard and the graphics card is not necessary. In particularly compact or low-cost devices is no need for a dedicated graphics memory and instead of the main memory of the computer used concomitantly (see Unified Memory Architecture, Shared Memory ), which has a negative impact on performance.

Suppliers of IGPs:

• AMD
• Intel
• Nvidia
• PowerVR
• SiS
• VIA Technologies

Latest Laptops with PCIe interface can have a replaceable graphics chip (see Mobile PCI Express Module), a fact that did not (yet) enforced as standard.

Business Solutions

These are full-fledged graphics cards, where little attention is paid to the 3D features, but are designed to provide above all a sharp and high-contrast image. There are also variants with 3D optional functions, especially for CAD applications.

Games graphics cards

These cards are available in various price ranges from around 25 € up to 1000 €, with the expensive cards reflect what is technically feasible in 3D representation. In Games Card compete mainly AMD ( AMD Radeon series) and Nvidia ( GeForce series) with each other, whose chips are used by a variety of manufacturers on their graphics cards. There are also vendors such as S3 Graphics, Matrox ( one of the pioneers of 3D gaming graphics cards, but was pushed back by the powerful competitors in the professional market ) and XGI Technology, but play only a minor role and mostly in office PCs using find.

Since most games for Microsoft's Direct3D interface ( part of the Windows system component DirectX) are developed games graphics cards are optimized for maximum performance with this system. Graphics cards that offer full hardware support for the current version of DirectX can calculate virtually all technically feasible 3D - rendering capabilities in real time. Some games manufacturers rely on OpenGL but, above all, id Software.

Since 2006, the version 10 of DirectX is on the market, but that only works in conjunction with the Microsoft operating systems Windows Vista and Windows 7. DirectX 10 is supported by Nvidia from the Nvidia GeForce 8 - series and upwards, on the part of the AMD cards of the ATI Radeon HD 2000 series and up. Cards from the ATI Radeon HD 3000 series support even already the successor version of DirectX 10.1, which is shipped with the Service Pack 1 for Windows Vista and only minor amendments introduced (use in just a few games, such as the air combat game HAWX or Bethesda's Skyrim ). DirectX 10 is increased many limitations in shader programming and should have a lower overhead than DirectX 9, so that the running Direct3D commands to vonstattengehen faster. The disadvantage is that since the introduction of DirectX 10 only a few games are optimized for DirectX 10 ( the most prominent example: Crysis ), since the commercial sales of Windows Vista only on 30 January 2007 and began using the new DirectX 10 effects of a require enormous computing power and hence only on high- end graphics cards will work satisfactorily. Many new games support often still only DirectX 9 and sometimes parallel DirectX 11, DirectX 10, therefore, has only a minor importance.

Since late 2009, there are DirectX Version 11 This version is in the cards from ATI (or AMD ) from the " HD5000 " series and from the " GTX 400 " series from Nvidia support. The launch of DirectX 11 was better than the DirectX 10 because it at launch of Windows 7 and therefore DirectX 11 is already a game with DirectX 11 gave ( Battle Forge ) and more quickly followed. However, the games support all still DirectX 9, which makes it also runs on Windows XP.

In addition to Direct3D there as another graphics API OpenGL in the current version 4.2, which features a slightly more functionality than Direct3D 11.

Professional Solutions

These are especially graphics cards for CAD and GIS applications. The cards offer special for CAD / GIS necessary functions that can only be emulated on "normal" graphics cards and thus use much slower. After the last special chip supplier 3Dlabs in 2006 set the business, offer only AMD ( under the brand name ATI) and Nvidia solutions for OpenGL workstation segment on. The two companies are taking advantage of their derivatives game graphics card chips. These are then optimized using a modified ROM and drivers on the 2D representation of OpenGL and no longer on the 3D representation of DirectX and OpenGL. The drivers of these graphics cards support the drawing of several million lines smoothed and User Clip Planes. Although the hardware between games and OpenGL 3D chips chips differs only minimally, cost considerably more professional cards. The reason for this is to optimize the driver, the extensive customer service, the Workstation client must be stopped, and the very expensive SRAM, with some graphics cards are equipped. Furthermore, additional skills are often available as DisplayPort connectors to use a higher color gamut, or the projection of a large area with multiple image sources. The product lines are called AMD ATI FireGL or FirePro now AMD and Nvidia Quadro FX.

Others

In addition to the DirectX graphics cards described above, there are special cards that support only OpenGL. These are commonly used in animation and are now totally uninteresting for players, since most PC games only support DirectX ( however differently on the Macintosh platform). By default, each dominated today's DirectX graphics card and OpenGL, conversely, is not however the case.

Since the dawn of programmable graphics pipeline in 2000, it is possible the computing power of graphics processors for the calculation of parallelizable arithmetic operations, such as occur in technical and economic simulations, for example, to take advantage of. This application is referred to as GPGPU (General Purpose Computation on Graphics Processing Unit), also see in CUDA.

Software graphic interface

In order to use graphics cards to develop individually without hardware and software for each, there are various software graphic interface.

Especially interesting on basic functional level is the BIOS that important text and graphics output functions provides that are used, inter alia, of text consoles under DOS or Linux. These functions are relatively slow, but work reliably on each graphics card.

In most of today's operating systems is an abstraction layer between applications and hardware, the so-called device drivers. Without these programs need to address the hardware directly, but this would be due to the differences between graphics cards perform at a high level of specialization and thus a high programming effort for the support of many graphics cards. Since graphics card drivers can also offer very different functions, different graphics APIs have been developed over time to facilitate access to these functions. The best known among them are OpenGL, DirectX (specifically: DirectDraw, Direct3D ) and quartz, which allow the programmer to easily and independently of the graphics card to display 2D and 3D graphics. For DirectX and quartz, the interfaces do not necessarily imply hardware 3D capabilities of the graphics card, use this but if they are present. Older 3D applications can run even on computers with integrated graphics or a simple 3D map in principle, but relatively slowly and / or less visually appealing.

For the operating system eComStation the universal panoramic driver was developed, which serves all current graphics cards.

Software problems with graphics cards

Since many graphics cards nowadays allow the liquid watching videos by the computer through hardware acceleration and also have many graphics cards with a TV -out port, it is natural to connect the computer to a TV or VCR. However, it prevent some manufacturers with the video card driver or the video card itself, to combine both skills together. So it happens that when playing videos while the entire user interface is visible, the video itself is not. Under Linux, then (ie, the computer monitor ) works, for example, the XVideo implementation only when the primary display, but not the TV -out port. This problem can usually get around by turning off the hardware acceleration for video decoding, but the video is then often no longer liquid.

It is believed that such limitations are included to prevent the user of the recording of the video by a video recorder. Anyway, read that products from Macrovision ( a company that is known for a copy protection method ) were integrated into the graphics card in some appropriate handbooks.

A specific case is the fglrx driver from AMD, the (currently) does not support hardware-assisted video playback on the TV output.

Another problem was and still is the use of multiple VGA-compatible graphics card, as it may be the case in PCI systems. In this case, the operating system does not support any free combination, not even of graphics cards from the same manufacturer. By upgrading the ROM on the card can be created here but sometimes remedy.

Manufacturer

Manufacturer of graphics chips: 3dfx, 3Dlabs, AMD, Alliance Semiconductor, ARK Logic, ArtX, ATI Technologies, Avance Logic, Bitboys Oy, Chips & Technologies, Cirrus Logic, Matrox, NeoMagic, Number Nine, Nvidia, Oak Technology, Rendition, S3 Graphics, S3 Inc., SiS, Trident, Tseng Labs, Western Digital, XGI.

Manufacturer of graphics cards: ABIT, Albatron, AOpen, Asus, ATI Technologies, AXLE3D, Club 3D, Connect3D, Creative Labs/3DLabs, DFI, Diamond Multimedia, ELSA Technology, EVGA, Elitegroup, Gainward, Galaxy Microsystems Ltd., GeCube, Genoa. Gigabyte, Hercules Graphics, HIS, Leadtek, Matrox, MSI, miro, Number Nine, Orchid Technologies, Paradise Systems, pixel View, PNY PowerColor, Quantum3D, Sapphire, Sigma, Sparkle, SPEA, STB Systems, TerraTec, Video Logic, Video Seven, XFX, Palit, Zotac.

Today only AMD / ATI Technologies, Nvidia, Matrox and S3 Graphics as a graphic chip manufacturers in the market, as well as AMD, Intel, Nvidia, SiS and VIA Technologies is a manufacturer of integrated graphics solutions.