The Display Adapter

The Display Adapter

The display adapter in a computer arrangement ( frequently called a graphics card ) is the hardware responsible for generating the images that appear on the system ‘s ocular display unit. In the beginning IBM personal computers the display adapter took the form of an expansion card that was mounted in one of the Industry Standard Architecture ( ISA ) slots on the system motherboard. It used the Monochrome Display Adapter ( MDA ) standard introduced by IBM in 1981, and was lone capable of displaying text at a sieve resoluteness of 720 adam 350 pixels .
modern display adapters can take the form of a separate poster or may be integrated into the motherboard. They are able of operating at very eminent resolutions and displaying millions of color. even relatively humble display adapters can now handle calculator animations, television playback, and 3D synergistic games. indeed, it is probably the calculator gaming diligence that has been the main driver behind the development of high-performance display adapter technology. other crucial drivers include the about universal borrowing of the graphic user interface ( GUI ), a phenomenal growth in the numeral of multimedia applications, and the popularity of Internet streaming media .

An MDA video card
An MDA television tease

Displaying gamey resoluteness images and millions of color demands a considerable amount of both processing ability and computer memory, due to the enormous amount of information that must be processed. tied a relatively meek screen solution ( by today ‘s standards ) of 1024 x 768 pixels gives a total of 786,432 pixels. If a color depth of 32 bits per pixel is required, the amount of memory required to store a individual screen of data ( referred to as a skeleton ) is 786,432 ten 32 = 25,165,824 bits of data, or 3 megabytes.

The number of calculations that must be performed when rendering 3D scenes for many computer games would badly tax the capabilities of the calculator ‘s main processor. In order to run state-of-the-art calculator games and other computationally intensive graphic applications such as flight simulators and virtual medical training environments, a mighty graphics arranger is required. This will constantly be an accessory card with its own graphic march unit ( GPU ), and anything up to four gigabytes of consecrated memory. These high-end display adapters have a inclination to be bulky, consume a set of baron, and require their own cooling system .

An Asus ATI EAH5750 FORMULA PCI-E display adapter
An Asus ATI EAH5750 FORMULA PCI-E display arranger

The display arranger shown above is one of the mid-range products presently available and has a price tag somewhere between £100 and £160. It has a gigabyte of RAM, and can produce output in a range of signal formats including VGA ( at a maximal resolution of 2048 ten 1536 pixels ), DVI ( at a maximal resolution of 2560 ten 1600 pixels ) and HDMI ( the respective video standards and output formats will be described late on this page ) .
When considering the leverage of a graphics circuit board desirable for bet on or other diagrammatically intensifier applications, you will besides need to take several early factors into history. The organization motherboard must have an allow expansion slot for the card ( normally a PCI Express slot ), and you should check that there is enough space inside the case to physically accommodate the new tease. Make sure that the exponent provision has both sufficient capacity and the justly number and type of cable connections required. Operating arrangement compatibility may besides be an exit, since many cards require a minimum version count of DirectX in ordain to function correctly ( DirectX is a collection of application program interfaces ( APIs ) used in the program of games and other graphics-intensive applications ) .
The specifications for the Asus EAH5750 illustrated above state that a 500W or greater world power supply is needed, with two 6-pin PCI-E baron connectors. The host operating system should support DirectX 11, which efficaciously means the system must be running Windows Vista or above .
A expose arranger built into the motherboard ( much referred to as a graphics accountant ) normally has a small come of RAM that is augmented with some of the organization RAM ( the RAM “ borrowed ” from the system becomes unavailable for use by the organization ). This on-board graphics capability is constantly incapable of coping with the demands of sophisticate 3D calculator games, but is perfectly adequate for function type applications, surfing the network, and most clientele software .
The majority of new computers have an on-board display arranger, although the on-board graphics can be disabled if a consecrated graphics card is installed, either by the user explicitly disabling the on-board accountant in BIOS or mechanically when the organization detects the presence of an extra graphics card. note that for little shape factor computers it may be necessity to beginning a desirable card from the manufacturer of the arrangement, assuming one is available. For laptop and netbook computers, upgrading the display adapter is not normally an choice .

Screen resolution, colour depth and memory

The images you see on your calculator proctor are created using millions of bantam dots called pixels, arranged in rows and column on the screen. information about each pixel to be displayed is stored in video memory. This data is read by the display arranger and converted into a signal that the monitor hardware can use to display the pixel. The color of each pixel is determined by the volume of three primary colours – crimson, green and blue sky .
The signal send to the monitor must tell it how much bolshevik, park and blue output to produce for each pixel. These red, green and blue values are often referred to as RGB values. The sum number of pixels that make up the image will depend on the width of the image in pixels ( its horizontal resolution ) and stature of the image in pixels ( its vertical resolution ), and can be calculated by multiplying these values together. Thus the total number of pixels for a screen door resolution of 640 x 480 pixels is equal to 640 x 480 = 307,200 .
screen solution is important because it determines the sum of detail that can be shown in an prototype. The greater the number of pixels displayed, the more information can be conveyed to the drug user. In the very low resolutions used on early calculator systems, it was possible to pick out person pixels on the screen quite well. With today ‘s high resolution displays this is rarely possible .
The leftover hand image downstairs is 320 pixels wide by 200 pixels high, which was the highest color graphics resolution available for the IBM personal computer in 1981 ( eminence that neither of the images shown could have been displayed on a CGA monitor, as the CGA display arranger could only display four color ). The proper hand image is a small part of the first image magnified eight times. You can clearly see the individual pixels that make up the trope .
The physical size of the display filmdom has changed little in real terms. The original IBM 5153 CGA color monitor had a 12 column inch screen as opposed to the 15 – 19 column inch screens in widespread consumption today. In contrast, display resolutions now have several times as many pixels in both the horizontal and vertical dimensions. The vogue has therefore been for pixel density to increase, giving sharper and more detailed images. There is of course a cost, as the amount of memory required to store the picture data has increase dramatically .

display adapter 03 display adapter 04
A 320 x 200 pixel image ( left ) and a close-up position ( right field )

The phone number of pixels entirely is not the only factor that determines how a lot memory is required to store a individual screen double. For a simpleton black and white image in which each pixel is either black or white ( or, as in early monochrome displays, amber or k ), merely one spot is required to store information about a individual pixel, since it is either clear or dark, on or off .
The count of bits required to store data about the colours used in a detail picture is known as its color depth, expressed as bits per pixel. The identical beginning display adapters able of displaying coloring material graphics were limited to a fix palette of four colours. The color to be displayed could be determined using merely two bits, which allowed four distinct coloring material values ( 0 to 3 ). The count of colours available soon rose to sixteen, requiring four bits of data for each pixel .
When the number of bits per pixel rose to eight a total of 256 colours became available, allowing reasonably good quality images to be displayed. obviously the sum of memory required to store an picture at a color depth of 8 bits per pixel is far greater ( eight times a much, in fact ) than the amount required for a monochromatic image .
The number of bits per pixel quickly doubled to sixteen, allowing a total of 65,536 freestanding colours to be displayed on the screen ( 216 = 65,536 ). When a far eight bits were added, giving a color depth of 24 bits per pixel, a stagger 16,777,216 ( 224 ) separate colours were available. This is far greater than the sum number of pixels that can actually be displayed by any computer monitor presently available .
Because of the way that video memory is accessed in most cases, the number of bits used to store information about a single pixel is actually 32, and most operating systems and applications refer to a 32-bit color depth. The amount of memory required to store a one riddle trope in bytes is calculated as follows :
memory ( in bytes ) = width ( in pixels ) x acme ( in pixels ) x bits per pixel ÷ 8
so, for a 32 snatch picture with a resolution of 1024 x 768 pixels, the calculation would be :
memory = 1024 ten 768 adam 32 ÷ 8 = 3,145,728 bytes ( or 3 megabytes )
As mentioned earlier, the color displayed on screen for a given pixel depends on the intensity of the three primary colours – red, green and blue. For a color depth of 24 bits per pixel, the volume of each color is stored as 8 bits of information, giving a total of 256 possible intensity values for each of the three elementary colours. This yields a total number of color permutations of 16,777,216 ( 256 adam 256 x 256 ) .
For a color depth of 16 bits per pixel, the numbers do n’t break down therefore neatly. A compromise solution is used in which 5 bits are used for each of the crimson and blasphemous semblance components ( 32 potential intensities ), and 6 bits for greens ( 64 potential intensities ). The sum number of color permutations here is 65,536 ( 32 ten 32 ten 64 ). many people find it unmanageable to tell the deviation between 16-bit and 24-bit color images, and for most business applications a color depth of 16 bits per pixel is absolutely adequate. The same image is displayed below at versatile color depths .

24 bits per pixel (17,639 colours used)
24 bits per pixel ( 17,639 colours used )

8 bits per pixel (245 colours used)
8 bits per pixel ( 245 colours used )

4 bits per pixel (16 colours used)
4 bits per pixel ( 16 colours used )

1 bit per pixel (2 colours used)
1 bit per pixel ( 2 colours used )

Of the above images, only at very low color depth ( 4 bits and 1 bit per pixel ) is there an obvious abasement in quality, although closer inspection of the 8 bits per pixel version will besides reveal some shortcomings. even so, a reasonable image quality is achieved at 8 bits per pixel .
In commit, although some applications can take manipulate of the screen door and change the coloring material depth used for displaying the screen images they produce ( some games for exercise ), the function organization will determine the tinge depth used by the display adapter. By default, the operate system normally selects the highest color depth available ( normally 32 bits per pixel ) in order to ensure the best quality display possible, although the exploiter may select a different discolor depth if they wish .
You may note from the images shown above that tied using only 16 colours, the image manages to achieve a fair approximation to the 24-bit translation. This would not be the encase if the 16 colours available were restricted to entirely those present in the 16-colour palettes provided by the CGA and EGA display adapters ( see below ) used on early IBM personal computers .
In fact, the sixteen colors in the pallette can be chosen from the entire range of 16,777,216 colours, because the intensities of each of the RGB values used for the semblance in the palette are stored using 8 bits. The 4 bits of information stored for each pixel in the screen prototype is basically an index issue that indicates which of the sixteen color is to be used for that pixel .
The area of video memory used to store the pixel information for a unmarried shield trope is called the frame buff, and as we have seen its size will depend on the resolution and semblance astuteness used. The area of conventional memory reserved for television RAM is simply besides modest to hold a screen image for high resoluteness 32-bit images, so virtually all modern graphics cards come with their own video memory .
Most on-board graphics controllers are capable of handling the requirements of 2D applications, although they normally “ adopt ” some of the system ‘s main memory in decree to store persona data, vitamin a well as the results of any picture work operations they may undertake. Display adapters adequate to of running today ‘s 3D synergistic games need big amounts of memory, both for storing screen images and for storing the results of the big number of calculations required for creating each new trope. The memory used is much faster than standard system memory, and we will be looking at the development of video RAM by and by on this page .

Display adapter evolution

As mentioned above, the first base IBM personal computer had an MDA expose adapter that could only display monochromatic text. The end product consisted of 80 columns x 25 rows of characters, with each fictional character having on-screen dimensions of 9 pixels wide ten 14 pixels gamey. From this information, it is potential to infer a screen settlement of 720 ten 350 pixels ( an aspect proportion of 72:35 ) and a tinge astuteness of one bit per pixel. It should be noted however that individual pixels could not be directly manipulated .
The MDA card had 4 kilobytes of television RAM on board, and a screen refresh rate of 50 Hz ( i.e. the contents of the filmdom were redrawn by the expose hardware fifty dollar bill times every second ). The characters displayed were taken from the 256-character IBM character set which was stored in the display adapter ‘s read-only memory ( ROM ). It was however possible to display graphics of a screen and evening to create vestigial personal computer games through the imaginative use of some of the more alien characters in the IBM character plant .
The illustration below shows the claim sieve of a very early Microsoft game designed for use with the MDA display adapter. The IBM logo appears to have been created by repeatedly displaying an IBM pulley character at the necessity screen coordinates .

The title screen for an early Microsoft game designed for an MDA display
The title screen for an early Microsoft game designed for an MDA display

The arranger itself was equipped with 4 kilobytes of memory, and utilised a dedicated area of the independent system memory ( known as video RAM ) starting at memory address B000:0000. Each blind character was stored in memory as a two-byte password ( a word is the amount of data that can be stored by a processor ‘s registers, in this case 16 bits ). One byte contained the fictional character code, while the early contained the attribute ( e.g. high intensity, underscore, blinking etc. )
The contents of the screen were efficaciously held in a planar array of words in which each array chemical element corresponded to a particular 9 ten 14 pixel matrix at a particular character status on the shield. The display adapter would read the contents of the video recording RAM many times per moment to determine what fictional character elements to display at each character put. Changing the respect of a especial align element in memory resulted in a change in the character displayed at the corresponding screen coordinates ( this is referred to as memory mapped I/O ) .
1981 besides saw the arrival of IBM ‘s Colour Graphics Adapter ( CGA ), which featured 16 kilobytes of on-board video RAM. The CGA card occupied an ISA slot on the system ‘s motherboard. It had a screen refresh pace of 60 Hz, utilized system memory starting at address B800:0000, and supported several clear-cut modes of operation. The highest resolution graphics mode that could be achieved was a 640 adam 200 pixel black and white display. This is a lower screen resolution than MDA ‘s 720 ten 350 black and white display, although individual pixels could be manipulated programmatically .
Using the like resolution, CGA could display 80 columns x 25 rows of characters in 16 colours ( requiring a color astuteness of 4 bits per pixel ), although the consumption of an 8 adam 8 pixel matrix for each fictional character resulted in textbook of significantly poorer choice that that produced by an MDA batting order. Like MDA, the CGA menu featured a 9-pin D-sub ( DE-9 ) connection. It could be connected to a television receiver, or to a dedicated monitor capable of displaying color in order to take advantage of the color output. Note however that IBM did not actually produce a color monitor until 1983, when they introduced the 5153 Personal Computer Color Display .
As with MDA, fundamental graphics could be displayed at the 640 ten 200 ( or lower ) resolving power by cagey use of graphic characters, enabling all 16 colours in the CGA palette to be used. The color in the CGA palette were determined according to the RGBI discolor model, in which the first three bits represent the bearing ( bit is set ) or absence ( bit is not set ) of the loss, park and blue color components respectively, while the fourth spot increases the intensity of all three colour components if set ( if a morsel is set, it has a prize of 1, otherwise it has a respect of zero ) .
The highest solution color graphics mode accomplishable was 320 ten 200 pixels. In this manner, every pixel could be manipulated programmatically. The display was limited to precisely four of the 16 CGA colours, although two standard palettes were available in which the background color could be any one of the 16 CGA colours ( the default option background color in both cases was black ). The CGA color palettes are illustrated below .

The CGA colour palettes
The CGA discolor palettes

The high-intensity translation of Palette 0 or Palette 1 can be accessed by the platform setting the high-intensity bit. A third 4-colour palette dwell of cyan, loss and white plus the background color can be used if the complex color burst bit is disabled by the program, although it is not formally documented by IBM and can only be seen on RGB monitors .
many games were written using this palette because it was felt to be more attractive than the official palette options. Additional discolor could be approximated using dithering ( placing pixels of a unlike discolor next to one another to create the appearance of a third base color ) although at such depleted resolutions it was reasonably easy to pick out the coloring material of individual pixels. The images below are screenshots from early personal computer games running on a CGA display adapter. The first image demonstrates the function of dithering .

IBM's Castle Master, circa 1990 (Palette 0)
IBM ‘s Castle Master, circa 1990 ( Palette 0 )

IBM's Alley Cat, circa 1983 (Palette 1, high intensity)
IBM ‘s Alley Cat, circa 1983 ( Palette 1, high intensity )

Like MDA, CGA used the fictional character set stored in its onboard ROM when in text manner. The MDA cards were the democratic choice for business practice however, because the display textbook was of a much better choice than that produced by CGA cards, and the expose monitors required to take full advantage of the capabilities of a CGA card were quite dearly-won. As an total bonus, IBM ‘s initial MDA offer included a twin printer interface in addition to the DE-9 connection .
ironically the arrival of the Enhanced Graphics Adapter ( EGA ) standard in 1984 actually increased the popularity of CGA cards, because their price fell importantly. As a result, PCs at the lower end of the grocery store produced by both IBM and their competitors tended to be equipped with CGA display adapters .
up until 1987, many computer games were however being written for entrance flush computers that featured a CGA card. 1987 saw the insertion of the new Video Graphics Array ( VGA ) standard, efficaciously making EGA the newly entry-level display standard .
EGA was introduced by IBM in 1984 for its newly PC-AT model personal calculator. Like its predecessors it was mounted in an ISA slot, featured a DE-9 connection, and had a screen refresh rate of 60 Hz. It besides displayed 80 columns x 25 rows of characters in text modality, but the maximum resolution was increased to 640 x 350 pixels ( an view proportion of 64:35 ) and more colours were available .
EGA cards initially included 64 kilobytes of on-board RAM, but this was increased to 256 kilobytes in subsequently versions. A 16 kilobyte ROM chip provided extra graphics functions not found in the system BIOS. In textbook mode, EGA used the lapp area of memory ( starting at address B800:0000 ) as EGA, but for its standard graphics modes it used memory starting at address A000:0000. Up to sixteen colours could be displayed at the same meter, a colour-depth of 4 bits per pixel .
note however that the colours displayed could be selected from a standard pallette of 64 colours, which utilised two bits per pixel for each of the loss, green and blue sky color components. This allows one of four values ( 0, 1, 2 or 3 ) to be selected for each of the loss, green and blue pixel elements. The 64-colour EGA palette is illustrated below, together with the default 16-colour pallette. eminence that the colours have been identified using their equivalent hexadecimal RGB values. note besides that the default palette has the same semblance as the 16-colour CGA palette .

The EGA 64-colour palette and default 16-colour palette
The EGA 64-colour palette and default option 16-colour pallette

The two-bit binary star color values for each of the crimson, green and blue pixel components defined the intensity of each coloring material for a given pixel. A value of zero indicates that the tinge is turned off. A value of 1 represents one-third intensity, 2 represents two-thirds saturation, and 3 indicates full intensity .
EGA cards could emulate both the original CGA graphics modes and 16-colour versions of the 640 x 200 and 320 x 200 CGA graphics modes ( these were limited to the 16 colours in the CGA RGBI palette ). They could besides work with an MDA monitor if the switches on the wag were set up correctly, and were able to operate in both the 640 x 350 MDA black and white graphics and standard MDA text modes .
EGA was superseded in 1987 by the Video Graphics Array ( VGA ) standard, introduced by IBM with the new PS/2 model personal computer. It was to become the death widely accepted display adapter standard. virtually every display arranger produced since 1987 can trace its origins back to VGA, and most graphics cards support one or more of the original VGA modes .
VGA is probably most closely associated with its 640 ten 480 pixel graphics resolution, in which it can display 16 colours ( a coloring material depth of 4 bits per pixel ). This resoluteness gives an aspect ratio of 4:3 which has become the standard aspect ratio for personal computer screen displays. At a resoluteness of 320 ten 200 pixels, VGA can display 256 colours ( a tinge depth of 8 bits per pixel ). All of the color in the VGA palette can be selected from a much larger palette in which each color component is represented using 6 bits, allowing 64 possible values for each of the bolshevik, green and blue components of a pixel ( 64 x 64 x 64 = 262,144 ) .
A textbook manner with a sieve resolution of 720 adam 400 pixels was available, allowing 80 columns and 25 rows of text utilising a character matrix of 9 adam 16 pixels. text could be displayed with a choice of 8 background colours and 16 foreground colours. VGA could potentially achieve a maximum resolution of 800 ten 600 pixels, and a highest sieve refresh rate of 70 Hz ( the criterion refresh rate for the 640 x 480 pixel graphics manner was 60 Hz ) .

IBM's PS/2 computer, circa 1987
IBM ‘s PS/2 computer, circa 1987

VGA differed in respective ways from its predecessors. One major change was that the signals sent to the monitor were analogue quite than digital, allowing for a more accurate representation of color ( this of course mean that older monitors that relied on transistor-transistor logic ( TTL ) signalling would not work with VGA ). Output was via a 15-pin mini-D connection ( see below ) that became known as the VGA connection, calm the most normally use connection type for calculator display equipment despite the emergence of several other types in holocene years .
Another invention was the practice of a single application-specific integrated circuit ( ASIC ) that replaced most of the discrete logic chips found on the MDA, CGA and EGA cards. This not only drastically reduced the overall size of the printed circuit circuit board required for belated addition cards, but besides allowed the display arranger to be located on the personal computer ‘s motherboard providing it was accompanied by the other components necessity for it to function. These included timing, television memory ( of which 256 kilobytes was required to implement the full functionality of VGA ) and RAMDAC chips ( RAMDAC will be discussed subsequently on this page ). IBM ‘s first PS/2 PCs were provided with an onboard VGA display adapter .

The DE-15 VGA 15-pin connector socket
The DE-15 VGA 15-pin connection socket

VGA can emulate many of the expose modes of EGA, CGA and MDA with the extra benefit that although the CGA and EGA modes were however limited as to the act of colours they could display, those colours could be selected from the much larger range of colours available to VGA. Furthermore, VGA used different areas of television RAM ( the address space ranging from A000:0000 to B000 : FFFF ), depending on the mode selected. By default, VGA used a 64 kilobyte segment starting at address A000:0000 for both VGA and EGA-compatible graphics modes, a 32 kilobyte area starting at B000:0000 for monochromatic text mode, and a second gear 32 kilobyte area starting at address B0000:8000 for color text mode and CGA-compatible graphics mode .
The use of separate memory areas for different display modes made it possible for two monitors to be connected to a one personal computer with both an MDA batting order and a VGA menu ( or EGA or CGA card ) installed. In the early 1980s this arrangement was typically used to display text-based information ( spreadsheet data, for case ) on an MDA expose while viewing associate graphic information ( such as charts and graph ) on a second, low-resolution CGA display .
A count of other display standards emerged during the 1980s, including the Hercules Graphics Card ( HGC ) which was a popular alternative to IBM ‘s MDA technology and achieved a resolution of 720 ten 348 pixels. ampere well as the standard text modes, the HCG was adequate to of displaying black and white graphics – something the MDA display was not able to do .
Software packages like the popular Lotus 1-2-3 spreadsheet broadcast provided support for the HGC, which as a consequence could display graph and charts generated by the platform adenine good as standard text-based spreadsheet information. IBM was the dominant force throughout most of the 1980s however, producing modern standards such as 8514/A and MCGA, both of which were released in 1987 ( the like year as VGA ) .
The 8514/A display arranger provided a higher resolution colour graphics modality ( 1024 x 768 pixels ) with a color astuteness of 8 bits per pixel, allowing 256 classify colours to be displayed at the like time. This higher solution only worked on intertwine monitors at a screen refresh rate of 43.5 Hz, however. An fretted monitor displays alternate lines on each refresh cycle in orderliness to make more efficient use of bandwidth .
The lock monitors proved unpopular because they tended to produce waver, and even at higher frequencies could cause eye tenor after prolonged consumption. however, the 8514/A display adapter was the beginning mass-produce display card to feature hardware acceleration ( see below ), albeit of a slightly express nature. This mean that it could undertake some of the display-related tasks normally carried out by the CPU. It was besides the precursor to IBM ‘s Extended Graphics Array ( XGA ) .
The Multicolor Graphics Adapter ( MCGA ) was offered by IBM as a cheaper alternative to VGA on some of its PS/2 model personal computers. While capable of 256 color graphics at a resolving power of 320 ten 200 pixels, MCGA ‘s 640 ten 480 resolution was restricted to a black and white display lone .
later PS/2 model computers featured IBM ‘s XGA display arranger, first base introduced in 1990. Like the 8514/A standard, XGA featured hardware acceleration. The initial version provided resolutions of 640 ten 480 and 800 x 600 pixels at a color astuteness of 16 bits per pixel, allowing 65,536 discriminate colours to be displayed ( this discolor depth was referred to angstrom high tinge ) .
XGA besides supported a resolution of 1024 x 768 pixels with a semblance depth of 8 bits per pixel ( 256 colours ). XGA-2 increased the color depth for the 1024 x 768 resolution to 16 bits per pixel ( high color ), and added accompaniment for a 640 x 480 resolution that could display 16,777,216 colours, a color depth of 24 bits per pixel ( referred to as true color ) .
From the end of the 1980s onwards, IBM faced increasing competition from its competitors in the race to develop bigger and better display adapters, and effectively relinquished its character of defining video display standards. A new generation of graphics cards emerged that were loosely described as Super VGA ( SVGA ), a label effectively applied to a whole compass of proprietorship technologies that extended the capabilities of VGA .
There was no actual consensus among different manufacturers about what actually constituted SVGA, resulting in the proliferation of multiple standards, chipsets, resolutions, discolor depths and review rates. As a direct result of this, about every new display adapter produced required its own driver software to enable it to work with the computer ‘s manoeuver system. fortunately, most of these newly products could besides operate in standard VGA mode .
In an undertake to rationalise computer video recording graphics standards a number of manufacturers, led by the Japanese IT company NEC, established the Video Electronics Standards Association ( VESA ) in 1989. The initial purpose was to create a standard for 800 ten 600 pixel SVGA expose adapters. VESA has subsequently issued a total of graphics-related standards, and is even active in the development of raw standards, although the organization has been heavily criticised for a bankruptcy to disclose information about some older standards, and for levying brawny charges to non-members for software documentation relating to new standards ( many other related diligence standards bodies make their standards software documentation freely available for download ) .
The VESA criterion pertaining to SVGA is called the VESA BIOS Extension ( VBE ), and is a largely successful attempt to establish a standardized application programming interface ( API ) between application software and the display adapter. The most late adaptation of VBE is interpretation 2.0. Graphics cards that support the VESA SVGA standards ( either directly in hardware or via driver software ) allow programmers to take advantage of the standard API without having to write device-specific program code .
Since the arrival of SVGA at the end of the 1980s there have been a number of new display adapter standards, many based on IBM ‘s XGA standard, that reflect the increasing demand for high solution computer graphics and an ever more divers crop of drug user applications. The 4:3 and ( to a lesser extent ) 5:4 expression ratios that were omnipresent anterior to about 2003 have in many instances been displaced by widescreen formats on both laptop computers and standalone LCD monitors. These widescreen formats are democratic for a number of reasons, including the ability to display two pages of a document side by side on the sieve, and for viewing widescreen movies .
Some of the more significant standards are described in the table below ( note that all of these standards support a color depth of 32 bits per pixel ). The follow illustration gives an idea of how display resolutions have evolved.

The evolution of personal computer display resolutions
The evolution of personal calculator display resolutions

Current Display Adapter Standards
Standard Description
WXGA Widescreen Extended Graphics Array – an extension of XGA found on some recent laptop computers that provides pixel resolutions of 1280 x 800, 1366 x 768, and 1300 x 768. Note that the first two resolutions give aspect ratios of 16:10 and 16:9 respectively. The third resolution is a slightly narrower aspect ratio also found on some 26″ LCD television screens.
SXGA Super XGA – sometimes referred to as Extended VGA (XVGA). SXGA has a resolution of 1280 x 1024 and an aspect ratio of 5:4, the native resolution and aspect ratio for many 17″ and 19″ LCD monitors.
SXGA+ Super XGA Plus – most often found on 14″ and 15″ notebook computers, SXGA has a resolution of 1400 x 1050 (an aspect ratio of 4:3).
WXGA+ (or WSXGA) Widescreen Extended Graphics Array Plus – a non-standard set of display resolutions that usually refers to a resolution of 1440 x 900 (an aspect ratio of 16:10) found in notebook computers and many 19″ widescreen LCD monitors.
UXGA Ultra XGA – this standard has a resolution of 1600 x 1200 (an aspect ratio of 4:3), the native resolution for many 20″ and 23″ LCD monitors.
WSXGA Widescreen SXGA – at 1680 x 1050 (an aspect ratio of 16:10), WSXGA provides the native resolution for many 22″ monitors.
WUXGA Widescreen UXGA – this standard has a resolution of 1920 x 1200 (an aspect ratio of 16:10) found in a number of high-end notebook computers and many 23″ – 27″ widescreen LCD monitors.
QWXGA Quad Wide Extended Graphics Array – both Samsung and Dell have produced 23″ LCD monitors that support QWXGA’s 2048 x 1152 resolution (an aspect ratio of 16:9).
WQXGA Widescreen Quad Extended Graphics Array – at 2560 x 1600 (an aspect ratio of 16:10), this is the native resolution for a number of 30″ widescreen LCD monitors.

The graphics processing unit (GPU)

The graphics processing unit of measurement ( GPU ) is a dedicate microprocessor that performs most of the calculations required for drawing images on the screen. In the early days of personal computers, these calculations were carried out by the computer ‘s central serve unit of measurement ( CPU ) and the results were used to send image data to the frame fender .
The independent job of the display arranger hardware was to translate the information in the inning buff into a signal that the monitor could understand and display as an prototype on the blind. While this worked well for command line operating systems like DOS and moo resoluteness graphic modes that were limited to entirely a few discolor, the advent of higher resolutions and color depths together with operate systems and applications that featured graphic user interfaces ( GUIs ) meant that the resources of the CPU were being badly stretched .
In response, manufacturers of display arranger hardware began to produce cards that included specialised hardware designed to take over some of the process involved in producing double data. such cards were known by the generic condition accelerators, because they could carry out the needed tasks more cursorily than the central processing unit. rather of making all of the necessity calculations required to draw a window on screen, for model, the CPU could merely send a basic specification for the window to the graphics card, and the on-board acceleration hardware would perform the necessity calculations. The CPU was therefore able to carry out its other tasks more efficiently .
nowadays, the display adapter has its own processor that is dedicated to carrying out the enormous number of floating-point calculations necessary to render senior high school definition 2D and 3D graphics. The GPU on a modern graphics card must carry out a huge number of calculations in club to translate the cubic effigy data typically generated by gaming software and other 3D applications into a planar picture that can be displayed on the sieve. Due to the dynamic nature of such software, a large number of different images must be displayed on the screen every second to create an effective 3D experience for the user. Most GPUs are nowadays manufactured by ATI Technologies ( immediately a auxiliary of AMD ) or NVIDIA .

The ATI Cedar GPU in a Radeon HD5450 graphics card
The ATI Cedar GPU in a Radeon HD5450 graphics circuit board

The ATI Cedar GPU ( illustrated above ) is one of the less brawny graphics processors in the ATI Evergreen class of GPUs first introduced in 2009. It is seen here in the ATI Radeon HD 5450 graphics card, which is relatively cheap at around £40. Nevertheless, it is probably twice a fast as any comparable integrated GPU, and while credibly not suitable for running state of the art computer games can carry out 104 billion floating point operations per second base. The Radeon HD 5450 graphics tease will happily support up to three monitor displays simultaneously .
GPUs nowadays are highly twin, with literally hundreds of arithmetical and logic units ( ALUs ) and floating point units ( FPUs ). The necessitate for this degree of parallelism is due to the literally hundreds or thousands of polygons that make up a typical three dimensional scene in a modern computer game or simulation. The GPU must be able to perform calculations to determine whether each polygon is visible or not in a sieve image. It needs to process the adam, y and omega coordinates of each of the polygon ‘s vertices and convert them to x and y coordinates for a planar trope. It besides needs to determine the color of each pixel to be drawn to create the impression of a solid come on for the polygon. In a moral force 3D environment, this rendering process must be repeated many times per second .
The independent difference between a GPU and a conventional CPU is that the GPU is basically carrying out the like, specify act of operations on boastfully numbers of different data streams ( or pipelines ) simultaneously. The speed at which the GPU can achieve this current work is determined by the number of pipelines available ( which in turn is dependent on the number of floating point units introduce ), the travel rapidly of the GPUs clock, and the speed with which data can be transferred between the GPU and the expose arranger ‘s television memory ( this is normally many times faster than between the CPU and main arrangement memory ) .
The GPU will besides frequently include hardware that can perform video decoding functions, facilitating the playback of DVD or evening Blu-ray content, and the display of streaming television media. New technologies presently available for increasing the serve office available for graphic displays include NVIDIA ‘s scalable Link Interface ( SLI ) and ATI ‘s Crossfire, both of which allow multiple GPUs to be used to process a one image .
Although lacking the broad range of functionality characteristic of a conventional CPU, GPUs can perform float bespeak operations many times faster, and their pour process capabilities enable them to simultaneously execute boastfully numbers of small programs. For this reason, GPUs are becoming increasingly wide used in computer systems handling large-scale simulations, including a number of supercomputing applications .
The addition in the computational power of the display arranger has been matched by an increase in power consumption. This basically means that more heat will be produced by the display arranger, in finical by the GPU or video memory. many high-end graphics cards of the type used for dangerous calculator bet on now require their own cool systems .
As a minimum prerequisite, most cards will include a heat sink made of a heat-conducting metallic element that transfers heat aside from the core components. In many cases, a fan is required to maintain a sufficient stream of air over the heat sinkhole. In some systems, a heat sink is used that consists of a excavate chamber through which water is pumped in club to transfer heat away from the GPU. The use of a heat slump alone is referred to as passive cooling, and may well be sufficient for most integrate graphic restrainer chips .

A heat sink provides passive cooling a Radeon HD5450 graphics card
A heat dip provides passive cooling a Radeon HD5450 graphics poster

A water-cooled NVIDIA GeForce 9800 GT graphics card
A water-cooled NVIDIA GeForce 9800 GT graphics tease

Video memory

By the year 2000, about twenty years after the IBM PC beginning appeared, a typical high-end display arranger had in the order of 128 megabytes of video memory. By 2005, the first PCI Express graphics cards with 512 megabytes of video memory had appeared. The amount of video memory available is important because it determines how much information the graphics card can store about the images it is required to display .
The type of memory used for expose adapters was originally the same Dynamic Ram ( DRAM ) as that used for the independent system memory, but the increasing demands of calculator games and early graphically intensifier applications required a kind of memory that could be accessed much more quickly due to the measure of data used for each blind image and the large number of frames displayed per moment .
During the late 1980s and early 1990s, the type of memory most normally used for the image skeleton buffer was a dual-ported type of DRAM called Video RAM ( VRAM ). Dual-ported memory allows read and write operations to occur simultaneously, speeding up overall access times .
DRAM stores each bit in memory as a accusation stored on a microscopically humble capacitor. Because this load leaks away identical promptly, the contents of memory want to be refreshed after each read or spell operation. During the refresh cycle, no data can be read from or written to memory. Data words ( a word is the total of data that can be held in a processor register – typically two, four or eight bytes ) are stored in rows and column in memory. DRAM reads an entire row in a one learn operation, and discards everything except the datum bible located at the memory address specified for the read .
VRAM gets around the limitations of DRAM in two ways. First, it moves all of the datum words read in a one read operation into a branch rowing buff so that those bits can be accessed serially by the display circuitry, greatly reducing the overall phone number of take operations required. second, it has two input-output ports to allow data to be written to memory by the processor at the same time that it is being read by the expose circuitry. VRAM was more expensive than DRAM, but less so than former forms of dual-ported memory, making high-resolution, high focal ratio colour graphics economically viable. VRAM was finally superseded by synchronous DRAM ( SDRAM ) technologies in the late 1990s .
A number of different memory types have been used with display adapters since the 1990s. SDRAM is synchronised with the system bus, allowing the pipelining of instructions not possible with standard ( asynchronous ) DRAM, and improving memory access times. From 2003 onwards however, the vogue has been towards some form of double data rate synchronous dynamic random access memory ( DDR SDRAM ), first introduced in 2002. As the name suggests, this type of memory doubles the data rate by transferring data on both the heighten and falling edge of the clock bespeak generated by the memory busbar clock .
DDR SDRAM has a 64-bit wide data busbar width, so for the utmost clock frequency of 200 MHz ( and ignoring some non-standard manufacturer optimisations that achieved higher speeds ) the datum pace would be 200 ten 2 ten 64 = 25,600 megabits per second gear ( 3200 megabytes per second, assuming metric units are used ). Bear in beware that this is the maximal datum rate that can be achieved between the GPU and the video memory on the graphics card itself .
DDR2 SDRAM was introduced in 2003 and doubled the width of the memory data bus topology to 128 bits. The DDR2 specification besides increased memory bus speeds ( up to a maximal of 533 MHz ) but reduced the internal clock travel rapidly to half that of the memory bus, increasing data rates while achieving lower office consumption. Data is transferred at a rate of 128 bits for each half-cycle of the home clock, indeed for a bus frequency of 400 MHz, the maximum data rate would be 6400 megabytes per second ( 128 x 2 adam 400 x 0.5 / 8 = 6400 ) .
2003 besides saw the introduction of the fresh Graphics DRAM 3 ( GDDR3 ) memory standard for video recording memory developed jointly by ATI and the Joint Electron Devices Engineering Council ( JEDEC ). Based on DDR2 SDRAM, GDDR3 achieves a reduce power pulmonary tuberculosis that allows higher busbar speeds and reduces the complexity of cooling systems .
The successor to GDDR3 was GDDR4 but borrowing of this technology has not been far-flung. many current display adapters silent use GDDR3, with some graphics vendors opting for DDR3 SDRAM ( the successor to DDR2 SDRAM ) which increases utmost bus travel rapidly to 1066 MHz, doubly that of DDR2. The latest video RAM technology is GDDR5, which is based on DDR3 and provides significant improvements in bandwidth over GDDR3. The first products to use GDDR5 was ATI ‘s Radeon HD 4800 released in 2008 .

A Radeon HD 4870 graphics card and its Qimonda GDDR5 memory chip (circa 2008)
A Radeon HD 4870 graphics circuit board and its Qimonda GDDR5 memory chip ( circa 2008 )

Video memory technologies will no doubt continue to evolve at a rapid tempo, along with display arranger technology generally. Whilst the capacitance of the video memory is authoritative for storing both the image data and the intermediate results of the calculations required to populate the inning fender, bandwidth ( the total sum of memory that can be accessed in a given fourth dimension frame ) is besides a all-important gene in determining performance .
The bandwidth will be the factor that determines the maximal frame rate that can be achieved. note that skeletal system rate refers to the speed at which modern images can be generated and sent to the frame cushion. It should not be confused with refresh pace, which is the speed at which each screen image is re-written ( refreshed ) by the monitor .


RAMDAC stands for Random Access Memory Digital to Analogue Converter. The majority of monitors used with personal computers, from the insertion of VGA expose adapters in 1987 until the end of the 1990s, were cathode beam tube ( CRT ) devices that required analogue signals to drive the display. Computers are strictly digital devices, so naturally the picture data generated by the expose arranger and stored in the frame of reference buff is digital. Before it can be interpreted by the analogue proctor hardware, the digital data must be converted into an analogue format .
The RAMDAC reads the digital prototype data stored in the display adapter ‘s frame buffer, and converts it into analogue RGB signals. These signals determine the saturation of the red, green and aristocratic ( RGB ) components of each pixel. There is a separate digital-to-analogue converter ( DAC ) in the RAMDAC for each of the RGB components. Each DAC produces a different signal voltage, which is sent along the cable that connects the display adapter to the monitor .
The CRT proctor uses the RGB signals to control the operation of three electron guns ( one for each coloring material ). The electron guns traverse the screen horizontally and vertically in order to fire a stream of electrons at phosphor dots on the inner surface of the CRT screen, creating the color for each pixel. The gun traverse one row of pixels at a time, and the signals from the RAMDAC control the intensity of the beam fired by each electron grease-gun at each pixel. The entire operation occurs at a frequency of up to one hundred cycles per second .
Over the survive ten or so, CRT monitors have largely been replaced by flat filmdom LCD monitors that are less bulky and more exponent efficient. The change was gradual at first because the first flat screen products on the market were expensive, and in many cases the early offerings did not produce particularly good quality riddle displays by comparison with CRT monitors. For this cause, most display adapter manufacturers continued for some prison term to produce graphics cards that lone had the standard 9-pin VGA analogue output port, despite the fact that the modern LCD displays were digital devices. The LCD displays had to convert the entrance analogue signal bet on into a digital format in order to display the images correctly .
The amphetamine at which the RAMDAC can read the contents of the frame buffer zone and carry out the necessary digital-to-analogue conversion is a limit factor in terms of the freshen rates supported by the display adapter. many display adapters now offer both analogue and digital outputs, and some entirely provide a digital output. Likewise, most liquid crystal display monitors have both analogue and digital ports, with some only providing a digital interface ( the discipline of display arranger ports is cope with in more detail subsequently on this page ) .
once a separate component on the graphics card, the functionality of the RAMDAC is now normally integrated into the display arranger ‘s GPU. finally, there will be no prerequisite for any digital-to-analogue conversion and the RAMDAC will credibly disappear .

The Video BIOS

Like the system motherboard, the modern graphics calling card comes with its own basic input-output system ( BIOS ) chip that performs much the same function for the display adapter hardware that the main system BIOS performs for the arrangement as a wholly. It performs power-on self-test ( POST ) diagnostic functions such as testing the video memory .
Most modern graphics cards support the VESA BIOS Extensions ( VBE ) which provide a standard interface to the display adapter hardware for application software. VBE allows an application to detect the display modes supported by the graphics tease, and to set the mood required by the application. It besides allows the display adapter to communicate with monitors that support display power management sign ( DPMS ), enabling the monitor to be put into the appropriate power saving modality when not in use .
A display adapter seller or chipset manufacturer will much provide updates for video BIOS firmware that can be downloaded from their web site. As with all forms of firmware upgrade, it is crucial to make certain that the upgrade software ( normally called a flash utility ) precisely matches the video BIOS present on the display card. The seller ‘s upgrade instructions should be followed precisely, as failure to carry out a firmware promote correctly can render the circuit board inoperable. On some of the more recent graphics cards, the video BIOS has user-configurable settings that can be accessed through a special frame-up program .

The video system interface

The video arrangement interface defines how the expose arranger is connected to other internal system components such as the CPU and main organization memory. The sum of data being transferred over this interface has grown quickly, and has driven the exploitation of new and faster inner buses more than any other single factor .
many of the earliest video recording cards were connected to the calculator ‘s motherboard via an ISA expansion slot. The foremost ISA bus was featured on IBM ‘s XT personal computer and had a bus width of 8 bits ( i.e. it could transfer 8 bits of data in parallel ) and a clock frequency of 4.77 MHz. The IBM AT personal computer introduced an ISA bus with a width of 16 bits and a clock accelerate of 8.33 MHz .
IBM doubled the bus topology width again to 32 bits in 1987 with its proprietary micro-channel architecture ( MCA ), which was a feature of the new PS/2 personal calculator. Despite the fact that the MCA bus ran at a frequency of 10 MHz, it proved less popular than either ISA or the 32-bit 8.33 MHz extended ISA ( EISA ) interface that succeeded it in 1988 due to its proprietary nature and cost. Unlike MCA, the newly EISA interface was compatible with many of the older ISA cards. Despite some success in the waiter market for EISA-based phonograph record control and network cards however, EISA was not widely adopted for use in desktop personal computers. Both EISA and MCA soon passed into obscurity .
The early busbar standards mentioned above provided relatively low bandwidth that was normally shared between two or more attach devices. The concept of bus dominate was developed to allow a particular device to take dominance of the busbar for its own purposes, and to access the system memory directly. As the datum transmit rates required by display hardware increased, the bandwidth available on the ISA bus proved increasingly unable to meet the demand. consequently, the idea of a local busbar connecting one or more expansion slots immediately to the CPU was born .
The first base local anesthetic bus to be developed was the 32-bit VESA local bus ( VLB ) introduced in 1992, and clocked at 33 MHz. Although providing far higher bandwidth than ISA the VLB was not without problems, and its popularity waned after a couple of years as the peripheral part interface ( PCI ) started to appear on Pentium-based IBM personal computers and their clones from 1993 onwards. PCI resurrected the concept of a shared bus that implemented busbar overcome, and could compete with the VESA local bus in terms of performance .
PCI initially provided a bus width of 32 bits and a bus frequency of 33 MHz, although bus width was former increased to 64 bits with the PCI-Extended ( PCI-X ) interface, which supported bus speeds of up to 133 MHz. It solved many of the problems that dogged the VESA local anesthetic bus, and was the first interface to feature a hack and Play capability in which organization resources could be allocated to peripheral devices through software quite than by using manual frame-up procedures that involved setting jumpers or switches .
The first bus interface specifically dedicated to graphics data was the Accelerated Graphics Port ( AGP ) interface that appeared in 1997 on computers featuring Intel Pentium 4 or AMD Athlon XP processors. AGP provides a consecrated bus between the display adapter hardware and the memory accountant hub ( or Northbridge ) on the computer system ‘s motherboard. Unlike PCI, the share nature of which dictates that attached devices run at some preset divide of the front english bus ( the bus connecting the Northbridge to the CPU ), the AGP interface can theoretically offer the lapp bandwidth as the front side bus .
AGP translation 1.0 initially provided a 32-bit bus with a clock frequency of 66 MHz, and could transfer 32 bits of data per clock hertz. This mode was known as 1x because it provided the same utmost data transfer rate ( 267 megabytes per second ) as a 64-bit, 33 MHz PCI-X bus. AGP version 1.0 besides supported a 2x mode that doubled the throughput to 64 bits of data per clock cycle for a datum throughput of 533 megabytes per second .
AGP version 2.0 display adapters used a lower electric potential, required a longer slot on the organization motherboard, and supported 1x, 2x and 4x ( 1,067 megabytes per second ) modes. AGP 3.0 is the most late and effectively the concluding interpretation of AGP, and supports 4x and 8x ( 2,133 megabytes per irregular ) modes. Motherboards supporting the AGP interface have gradually been phased out over the final few years in favor of the new PCI Express interface .

An AGP 3.0 display adapter slot
An AGP 3.0 display arranger time slot

PCI Express ( PCI-E ) beginning appeared in 2004 and is the most holocene standard for display adapter cards at the clock time of writing. many devices presently on the commercialize support the version 2.x specifications ( the specification for translation 3.0 was published in November 2010 ). Newer versions of the standard are designed to be backwardly compatible with previous versions, and graphics cards designed for PCI-E 1.x should work on motherboards that provide PCI-E 2.x slots .
PCI-E differs basically from the original PCI in that it provides dedicate point-to-point serial links between two devices, rather than sharing the bandwidth of a single latitude interface. A single PCI-E radio link is a full-duplex ( bipartite ), four-wire duct known as a lane. Data is transmitted in packets, each of which includes the address of its destination node, over a packet switched network .
The network consists of a series of point-to-point connections linking devices to each early, the CPU and system RAM via a interchange ( called the root complex ) built into the motherboard. The slowest kind of connect on the network is a 1x yoke that consists of a single lane. Standard PCI-E connect designations are 1x, 2x, 4x, 8x, 16x and 32x ( the phone number in front of the adam represents the numeral of lanes used ) .
Motherboards presently on the market typically provide one PCI Express 16x slot and one PCI Express 1x slot, american samoa good as a match of standard PCI slots. Most graphics cards presently on the market require a PCI-E 16x slot. note that the size of a PCI-E slot will depend on its designation ( 1x, 4x, 16x etc. ), and whilst a PCI-E device designed for a 1x slot will normally fit into and work in a larger slot, the overrule is not broadly true. eminence besides that because the PCI-E root building complex is bridged to the PCI bus, you should besides ( in hypothesis ) be able to plug older PCI devices into a PCI Express slot, if it is long enough .
Some high-end motherboards offer two or more PCI-E 16x slots, allowing multiple display adapters to be installed. Some high-end graphics cards can support up to three monitors each, thus in theory two high-end graphics cards installed on the lapp motherboard would enable a computer to output different displays to six monitors simultaneously !
The datum rates supported by PCI-E were from the beginning much faster than those previously achieved. PCI-E version 1.x operates at a bus frequency of 2.5 GHz and provides a data rate of 250 megabytes per second per lane. PCI-E version 2.x doubles the bus frequency to 5 GHz and the datum rate per lane to 500 megabytes per second, giving an aggregate data pace for a 16x link of 8 gigabytes per second .

This motherboard features two 16x and two 1x PCI-E slots, and two PCI slots (far left)
This motherboard features two 16x and two 1x PCI-E slots, and two PCI slots ( army for the liberation of rwanda left )

Drivers and software

Driver software provided by the display adapter seller or chipset manufacturer provides the interface between the display arranger and the calculator ‘s operate system. A newfangled graphics card purchased for a system ascent will constantly be accompanied by a cd-rom hold driver software, utilities and tools. The pack software may well include double manipulation software as a value-added feature .
Thanks to the rapid travel rapidly of developments in the worldly concern of graphics hardware and software, the driver provided is much out of date by the clock time the tease is sold, so a visit to the chipset manufacturer or seller ‘s web site may be required in order to retrieve the most up-to-date versions of drivers and utilities. note that an manoeuver system seller will normally provide generic expose arranger drivers for standard VGA display modes and may provide their own drivers for specific hardware products, but it is preferable to use the manufacturer ‘s own driver software .
Some applications, particularly games, will alone function correctly with the latest drivers. For laptop and netbook computers that tend to have a highly customised hardware shape, it may be necessary to obtain drivers from the computer manufacturer .
While the driver software provides a basic interface between the operate on system and the display arranger hardware, many applications including games require an extra program interface in club to take advantage of the 3D capabilities of the graphics tease. This lotion programming interface ( API ) is provided by a standard library of programming functions such as Microsoft ‘s DirectX ( presently in version 11 ), which is actually a solicitation of APIs that handle versatile aspects of graphics programming .
many graphic applications written for the Windows operate system, including games, come with the ask version of DirectX. One of the features of DirectX is its ability, in some cases, to emulate missing hardware functionality in rate to enable application software to run even if the hardware itself lacks sealed necessitate features. Software that achieves this is said to have a hardware abstraction layer ( HAL ). Clearly however, performance in such cases is not going to be ampere good as it would be if the functionality were in full implemented in hardware .
silicon Graphics have developed their own cross-platform standard API, called the Open Graphics Library ( or OpenGL ) that provides much the same kind of functionality for application programmers as DirectX. many stream display adapters feature GPUs that support both standards .

Cables and connections

For reasons already mentioned, producers of expose adapters and calculator monitors have continued to support the 9-pin D-sub ( DE-9 ) connectors that first appeared with IBM ‘s PS/2 personal computer, which required an analogue display unit of measurement. closely all expose adapters and computer monitors presently on the market nowadays provide digital ports, with many continuing to support VGA connections and some offering a image of connection types .
The most common type of digital interface presently provided is the digital ocular interface ( DVI ). DVI ports come in four independent flavours, as shown in the diagram below. The inaugural distinction to be drawn is between single connection and dual associate configurations. Single connection DVI connections have eighteen digital pin connections and can handle screen resolutions of up to 1,920 x 1,200 pixels, but many graphics cards can now output images at the much higher resolutions that can be displayed on ( typically ) 30 column inch monitors. More signal lines are needed for these higher resolutions, and these can be provided via DVI double connect connectors that have an extra six digital bowling pin connections .

DVI port configurations
DVI port configurations

A second eminence is between DVI-I connectors that provide both digital and analogue end product signals and DVI-D connectors that provide alone digital output. DVI-I connectors are characterised by an extra four pin connections to the bequeath of the main body of the connection as illustrated above. These pins carry analogue signals, allowing a DVI-I larboard to be connected to an analogue expose unit .
If a CRT or LCD monitor only has a VGA D-sub 9-pin connection, the joining can be made using a desirable arranger cable. many graphics cards presently on the commercialize have either one or two DVI ports, equally well as a VGA D-sub port. It is frequently possible to connect two expose units to a one display arranger in order to split the display output across the two screens .
A new digital interface called DisplayPort was introduced in 2006 that is intended to finally replace both VGA and DVI connections between computers and their display units. DisplayPort has started to appear on a little act of graphics cards, monitors, and laptop computers, and is forecast to increase its market plowshare steadily over the following three to four years as support for VGA is phased out .
many high-end graphics cards provide ports to enable a computer to be connected to video recording and television equipment, vitamin a well as DVD players and game consoles. A 9-pin mini-DIN larboard can be used to allow analogue video recording remark and end product between the display arranger and an attached device via an appropriate cable television. This video in video out capability is frequently referred to using the acronym VIVO, and distinctive uses include being able to display computer-based media on a television sieve, or read television programmes on a calculator ( note however that some graphics cards that provide a 9-pin mini-DIN port merely support video recording end product ) .
A count of graphics cards are immediately besides providing a high definition media interface ( HDMI ) port to enable the computer to be linked to HD quick LCD television receiver sets .

A Radeon X850 XT graphics card with VGA, VIVO and DVI-I ports
A Radeon X850 XT graphics tease with VGA, VIVO and DVI-I ports

A Radeon HD4650 graphics card with VGA, HDMI and DVI-I ports
A Radeon HD4650 graphics menu with VGA, HDMI and DVI-I ports

Multiple display adapter technology

ATI ( now AMD ) and NVIDIA have developed competing technologies to allow two or more graphics cards to be installed on a single motherboard to create a single output with better graphics performance. The motherboard must of class support these technologies ( many motherboards now provide support for both ), and must provide the needed number and type of PCI Express slots .
The ATI technology is called CrossFire and first appeared in 2005. It initially required two Radeon x800, x850, x1800 or x1900 graphics cards, one of which had to be a CrossFire Edition interpretation ( the other had to be a standard translation ). The CrossFire Edition card acts as a master batting order, combining the outputs of the two cards together. The cards themselves had to be from the lapp intersection family, and were linked using a proprietary dongle that plugged into the primary DVI output port on each menu. The dongle provided a branch joining for the monitor cable television .
Later versions of the technology removed the want for a dongle, and for the presence of a CrossFire Edition master batting order. The two CrossFire enabled cards are hush linked together via a ribbon cable attached to connectors located at the lead of each menu. In some configurations, communication between the cards may take place rather via the PCI Express busbar provided allow driver software is available, although this has an impact on performance .
late developments include enhancements to the CrossFire engineering that allows it to support for up to four graphics cards mounted on the lapp motherboard. Since 2008, some dual-GPU cards can operate in CrossFire modality utilising an on-board PCI Express bridge, and two such cards can be linked together to provide a four-GPU CrossFire capability .

Two ATI Radeon graphics cards connected by a CrossFire dongle
Two ATI Radeon graphics cards connected by a CrossFire dongle

The NVIDIA engineering is called scalable Link Interface ( SLI ), and beginning appeared under that name in 2004. Like ATI ‘s CrossFire engineering, SLI initially required two ( preferably identical ) graphics cards to be installed in PCI Express 16x slots in a master-slave configuration. Motherboards that supported SLI provided a custom-make slot between the two PCI-E 16x slots into which a particular toddle poster was inserted .
The purpose of the paddle was to channel either 16 PCI-E lanes to the PCI-E slot holding the victor SLI batting order, or 8 lanes to each card ‘s slot, depending on which way round it was inserted. This was necessary because motherboards of the time merely did not have enough PCI-E lanes to provide 16 lanes to each slot ( this is not a problem on current motherboards ) .
The cards are interconnected by a unretentive printed circuit board called the SLI bridge that plugs into both SLI cards and allows data to be transferred directly between the cards. The unite output signal is sent to the monitor via the passkey SLI card. Two low-end NVIDIA cards can besides run in SLI mode without an SLI bridge by communicating via the PCI-E bus, since the bandwidth required is army for the liberation of rwanda less than that needed by high-end cards.

diverse configurations have emerged for graphics cards with double GPUs. Most of these configurations involve GPUs mounted on two separate printed tour boards ( PCBs ) that either have their own cooling units or a shared cool unit. The PCBs are connected by an SLI and a distinguish dedicated PCI-E 16x connect. The PCB holding the GPU that is designated as the maestro GPU is connected to a PCI-E 16x slot on the motherboard .
These dual-GPU cards have an on-board PCI-E bridge chip with 48 lanes that provide the same functionality as the media communications processor ( MCP ) that must normally be award on the motherboard in order for two graphics cards to be able to operate in SLI mode. This basically means that the cards can run in SLI mode on a motherboard that does not specifically support SLI. As with CrossFire, two dual-GPU SLI graphics cards can be paired on an SLI motherboard and bridged to provide a four-GPU capability .

Two NVIDIA graphics cards connected by an SLI bridge
Two NVIDIA graphics cards connected by an SLI bridge

informant :
Category : Tech

About admin

I am the owner of the website, my purpose is to bring all the most useful information to users.

Check Also


Manage participants in a zoom meeting webinar

Call the people who attend the meet as follows Alternate host host Who scheduled the …

Leave a Reply

Your email address will not be published.