What is the full form of a GP?
The Accelerated Graphics Port is an expansion bus standard designed for attaching a video card to a computer system to assist in the acceleration of 3D computer graphics.It was designed to be a successor to the PCI-type connections for video cards.Since 2004, the AGP interface has been phased out in favor of the PCI Express interface and only a few A GP models are available today.
The limits of a bus with shared bandwidth began to be pushed by successive generations of graphics adapters.A GP is a "bus" dedicated to graphics.
The AGP bus is a superset of the conventional PCI bus, and A GP cards must act as PCI cards.
The main advantage of AGP is that it provides a dedicated pathway between the slot and the processor rather than sharing the bus.The direct connection allows for higher clock speeds, as well as a lack of contention for the bus.
The address and data phases are separated in split transactions, which is the second major change.The host processes the address phases when the card is sent.The bus is not being used during read operations.
The handshaking of the bus is simplified.The total length is included in the request for A GP transfers, which are always a multiple of 8 byte long.Data is transferred in blocks of four clock cycles and pauses are allowed only between blocks.
The address and data buses are separated so the address phase doesn't use the main address/data lines.Adding an extra 8-bit "SideBand Address" bus over which the graphics controller can issue new AGP requests is how this is done.This results in increased data throughput.
This improvement in memory read performance makes it practical for an A GP card to read texture from system memory, while a graphics card has to copy it to the card's video memory.The graphics address remapping table apportions main memory as needed for texture storage.There is a maximum amount of system memory available to AGP.
The AGP slot first appeared on x86 compatible system boards.A flood of products followed from all the major system board vendors after the introduction of A GP support by Intel.[4]
The ALI Aladdin V. and the VIA Apollo VP3 were the first Socket 7 chipsets to support AGP.The first Socket 7 A GP system board based on the VIA Apollo VP3 was demonstrated by FIC in November 1997 and followed by the EPoX P55-VP3 which was the first to market.[5]
The Rendition Vérité V2200 was one of the first video chipsets to feature AGP support.Some early AGP boards used graphics processors built around PCI.This resulted in the cards not benefiting from the new bus, with the only improvement used being the 66 MHz bus clock.There were several examples of such cards.In order to exploit the new AGP feature set, Intel's i740 was designed to texture only from A GP memory, making it difficult to implement local board RAM.
A GP support was introduced into Windows 95 in the form of a patch.After applying the patch, the system became Windows 95 version 4.00.950 B.Windows NT 4.0 with Service Pack 3 was the first operating system to receive AGP support.The implementation of the A GPgart kernel module gave Linux support for enhanced fast data transfers.
"A GP specification 1.0" was released by Intel.3.3 V signals and 1 and 2 speeds were specified.1.5 V signaling, which could be used at 1, 2 and the additional 4 speed, was documented in Specification 2.0.The speeds are physically possible, but not specified.
The mandatory supports of extra register are only mentioned by Microsoft under the Universal Accelerated Graphics Port.The upgraded register include PCISTS, CAPPTR, NCAPID.There are new required register that include APBASELO.
An official extension for cards that required more electrical power, with a longer slot with additional pins.Professional computer-aided design applications used in the fields of architecture, machining, engineering, simulations, and similar fields were usually accelerated by A GP Pro cards.10
The final version of the standard did not include a 64-bit channel as an optional standard.
64-bit transfer for A GP8 is allowed by the standard.
A GP cards are compatible with limits."Universal" cards can fit into either type of slot, though 1.5 V-only cards won't go into 3.3 V slots.Both types of cards can be accepted in the unkeyed "Universal" slots.The 1.5 V portion of the card is used when plugged into an A GP Universal slot.The only cards that have keys for 1.5 V to prevent them from being installed in older mainboards are the GeForce 6 series and the Radeon X800 series.Some of the last modern cards with 3.3 V support were the Geforce 6 Series and 7 series.The Geforce 7300/7600/7800/7900/7950 cards are rare compared to their AGP 1.5v only versions.
Standard A GP cards can be used in a Pro slot.A 1.5 V or 3.3 V card can be used in either the AGP Pro or standard A GP configuration, or a Universal Agp Pro card.
Some cards have dual notches, and some motherboards have fully open slots, which can allow a card to be plugged into a slot that does not support the correct signaling voltage.The older 3.3 V cards have a 1.5 V key.
Some proprietary systems are incompatible with standard AGP, for example, Apple Power Macintosh computers with the Apple Display Connector (ADC) have an extra connector which delivers power to the attached display.Some cards may not work with others due to issues with the software.
The following comments were made by Mark Allen of Playtools.com.
Nobody makes AGP 3.0 cards.I can't find any manufacturers.Every video card I found was a universal 1.5V AGP 3.0 card.The universal 1.5V A GP 3.0 motherboards were the ones which claimed to be an AGP 3.0.It makes sense, because if anyone actually shipped a consumer oriented product, they would end up with a lot of confused customers and a support nightmare.You would have to be crazy to ship a product with less than 1 watt of power.
The power supplied by an AGP slot depends on the card used.The specifications for the various versions give the maximum current drawn from the rails.If maximum current is drawn from all supplies and all voltages are at their specified upper limits, an AGP 3.0 slot can supply up to 48.25 watt; this figure can be used to specify a power supply conservatively, but in practice a card is unlikely ever to.Many A GP cards had additional power connections to give them more power than the slot could provide.
Only a small number of new motherboards had A GP slots.Older chipsets with support for AGP were used in the production of the new motherboards.
Graphics processors of this period use a general-purpose standard that supports higher data transfer rates and full-duplex.An additional bridge-chip is required to convert PCIe signals to and from AGP signals.The additional board costs are due to the need for a separate circuit board and an additional bridge chip.
Several manufacturers of graphics cards continued to produce A GP cards.The first bridged cards were the graphics cards.There was a ceiling of the GeForce 7 Series in 2009.Club 3D, HIS, Sapphire, Jaton, Visiontek, Diamond, etc., all have AGP cards.The Radeon HD 2400, 3450 and 3650 were included.The HD 5000 A GP series was not available.There were a lot of problems with the Catalyst 11.2 - 11.6 AGP hotfix drivers under Windows 7.It's a workaround.Past versions of the A GP drivers are available from several vendors.
The last Windows release to support A GP was Windows 10 1511.The removal of A GP support was also considered.[20]
The same protocol is followed by the A GP bus after it is reset.The card has to act as a target and master.The "fast writes" extension allows the card to transfer data at higher speeds.
A GP transactions are allowed after the card is used.The card is the AGP master and the board the target.The card queue multiple requests correspond to the address phase, and the board schedules the corresponding data phases later.The card must be told the maximum number of outstanding AGP requests which can be queueed at a given time.
The starting address and length of AGP requests are always multiples of 8byte (64 bits) and they use a differentEncoding on command lines C/BE[3:0].The three low-order bits are used to communicate the length of the request.
The type of transfer to be performed next is indicated by three additional status bits.If the bits are 0xx, a previously queued AGP transaction's data is to be transferred; if the three bits aren't in use, the card may begin a PCI transaction or queue a request in-band using PIPE#.
Each A GP transaction begins with an address phase and a 4-bit command code.The possible commands are different.
High priority requests and long read commands were dropped by AGP 3.0.It mandated side-band addressing and dropped the dual address cycle, leaving only four request types: low-priority read, low priority write, flush and fence.
To queue a request in-band, the card must request the bus using the standard PCI REQ# signal.The card asserts the PIPE# signal while driving the AGP command, address, and length on the C/BE and AD lines.A dual address cycle is used if the address is 64 bits.Every cycle that PIPE# is asserted, the card sends another request without waiting for acknowledgement from the board, up to the configured maximum queue depth.The last cycle is marked by the deasserting of REQ#.
The PIPE# signal is not used if side-band addressing is supported.The signal is re-used for another purpose in the AGP 3.0 protocol.Instead, requests are broken into 16-bit pieces which are sent across the SBA bus.If the number of outstanding requests is within the configured maximum queue depth, a new request can be sent at any time.There are possible values.
Data transfers can be sent up to 8 the 66 MHz basic bus clock.The advantage of sideband addressing is that it mostly eliminates the need for turnaround cycles on the AD bus between transfers.
The data phase for a queue request will be performed next if the board indicates via the ST bits.Two priorities for each read and write, and each is processed in order.There is no limit on the number of low priority responses which can be delivered while the high priority request is being processed.
When the GNT# is asserted and the status bits have a value, a read response of the indicated priority is scheduled to be returned.At the next opportunity, the board will assert TRDY# and begin transferring the response to the oldest request in the indicated read queue.FRA ME#, DEVSEL# and IRDY# are some of the other bus signals that remain deasserted.Up to four clock cycles worth of data can be transferred without waiting for an acknowledgement from the card.If the response is longer than that, both the card and the board must indicate their ability to continue on the third cycle by asserting IRDY#.Wait states will be inserted if either one does not.At other times, the value of IRDY# is irrelevant.
During read responses, the C/BE# byte enable lines may be ignored, but they are held asserted by the board.
The RBF# signal indicates that the card is temporarily unable to receive more low priority read responses.No more low-priority read responses will be scheduled by the board.The card needs to be able to receive the end of the current response, the first four-cycle block, and any high-priority responses it has requested.
When GNT# is asserted and the status bits have a value, write data is scheduled to be sent across the bus.At the next opportunity, the card will assert IRDY# and begin transferring the data portion of the oldest request in the indicated write queue.If the data is longer than four clock cycles, the board will indicate its ability to continue by asserting TRDY# on the third cycle.Unlike reads, there is no provision for the card to delay the write, if it didn't have the data ready to send.
The C/BE# lines are used with write data, and may be used by the card to select which bytes should be written to memory.
The number of data transfers across the bus is indicated by the AGP 2, 4 and 8.Such transfers use a "strobe" signal from the data source.A GP 4 adds strobe signals.