Early PCs had their own hardware and all the computer components were integrated into the motherboard, leaving no scope for connecting external proprietary hardware. However, this concept was changed by IBM in its PC’s when it defined the open Architecture. The open Architecture introduced the concept of external buses and expansion slots, which allowed proprietary hardware to be connected to the PC. So the motherboard could be a minimal system, with only the supporting chips embedded into it, whereas, the expansion slots made it possible for external proprietary hardware to be connected.

Old generation Buses and slots

Here is a list of the old generation of buses and expansion slots not used in the motherboards anymore.

From the above chart it is clear that as more advanced microprocessors were developed, the individual buses had to be improved and widened. For example, with the advent of 80286 16 bits processor, the XT ISA 8-bit bus with 8-bit slots became quite inadequate and the new standard AT ISA had to be designed. The bus architecture was improved in order to take care of faster processors and faster peripherals devices.

New generation Buses

Here is a list of the new generation of buses frequently used in the modern motherboards.

PCI Bus

The proliferation of 32 bit CPUS and graphics intensive operating systems made it  obvious that the 8.33 Mhz AT ISA bus was no longer satisfactory. The PC industry began to develop alternative architectures for improved performance. Hence, the new architechture of  PCI bus was developed. The PCI is a 188 pin bus and has bus architecture that supports peripherals, such as hard disk drives, networks etc

The PCI architecture is capable of transferring data at 132 MB/sec – a great improvement over the 5 MB/sec transfer rate of the standard ISA bus. Another key advantage of the PCI bus is that it has automatic configuration capabilities for switchless / jumperless peripherals. It is possible to Auto – configure the bus.

PCI uses the linear burst technique of transferring data. It  is one of the unique aspects of the PCI bus because it will perform both burst reads and burst writes. This doubles the throughput compared to the buses without linear burst technique.

The devices designed to support PCI have low access latency reducing the time required for a peripheral to grant control of the bus after requesting access. The PCI bus supports bus mastering which allows one of a number of intelligent peripherals to take control of the bus to accelerate a high priority task.

PCI architecture also supports concurrency a technique that ensures that the microprocessor operates simultaneously with the masters, instead of waiting for them.

As an example computer parts like an Ethernet card and the Lan have control of the bus.

Finally PCI was developed as a dual voltage architecture.

In the PCI bus, to reduce the number of pins needed, the data and address bus are multiplexed together. It is the first bus to support the low voltage logic implementation of 3 volts. The clock signal provides timing for the PCI bus only and can be adjusted from dc o HZ to 33 MHZ Cache memory and JTAG support are also provided on the PCI bus.

AGP Bus

3D rendering is used in games and all types of presentation and CAD software. This  requires intensive processing and lots of memory. This lays strain on ordinary 3D video cards that make use of the PCI bus. The AGP (Accelerated Graphics Port) was developed by Intel and it uses a variation of the PCI bus slot. It provides a high speed data pathway between the 3D video card and main memory. It allows the AGP video card to utilize main memory for graphics work.

The AGP system renders the textures of 3D scene in a game or graphics application by using the main system memory rather than the video memory, thus allowing the graphics engine on the chip to perform faster rendering.

The AGP has deeply pipelined memory read and write operations which fully hide memory access latency.

Demultiplexing of address and data on the bus, allows 100% bus efficiency.

The 133 Mhz data transfer rate allows for real data transfers in the range of 500 Mb/sec.

The AGP is a system primarily meant for connecting fast 3D graphics video display devices, which makes gaming applications come to life.

The high speed AGP is totally independent of the PCI bus. It is a specialized bus system only meant for video display, whereas, other peripherals could make use of the PCI bus. For implementing AGP you need a motherboard of at least Pentium II ATX with a recent chipset that supports an AGP slot as well as an AGP video adapter. You also need a video driver for the particular operating system you are using.

PCI Express Bus (PCIe)

The PCI bus supports only 32 bits. The PCI standard has not been able to increase the width of the PCI bus. This has caused several bottlenecks in the development of new peripherals like 3D graphics cards. Even the AGP bus could not satisfy the need of serious gamers.  Hence, the need for faster and better bus was felt. This gave birth to the PCI –Express standard.

The PCI –Express bus is a new development and it is derived from the PCI bus. The PCI bus is a parallel connection bus, while the PCI-Express is a Serial connection bus and each device is connected to its own individual bus (called lane). PCI Express enhances the graphics capabilities of the video display. The PCI-Express is also known as the PCIe bus, and PCIe – x1, x2, x4, x6 etc and x16 are standards of PCIe bus.

The PCI – Express x16 can support two graphics cards. PCI –Express 16 is much better than the AGP bus for video rendering. The PCI – Express x16 is now mainly used in the high-end 3D video graphics cards developed by NVIDIA and ATI and is excellent for 3D texture rendering, etc, which requires large amounts of memory to function efficiently. Hence, for serious gaming needs the PCI-express x16 is the bus of choice, and it is better than the PCI and the AGP buses. Peripheral cards other than the video display using PCIe are very few as compared to the video display cards.