The PCI Express (PCIe) bus architecture (formerly 3GIO or 3rd Generation I/O) was introduced by Intel, in partnership with other leading companies, including IBM, Dell, Compaq, HP and Microsoft, with the intention that it will become the prevailing standard for PC I/O in the years to come.
PCI Express allows for larger bandwidth and higher scalability than the standard PCI 2.2 bus.
The standard PCI 2.2 bus is designed as a single parallel data bus through which all data is routed at a set rate. The bus shares the bandwidth between all connected devices, without the ability to prioritize between devices. The maximum bandwidth for this bus is 132MB/s, which has to be shared among all connected devices.
PCI Express consists of serial, point-to-point wired, individually clocked 'lanes', each lane consisting of two pairs of data lines that can carry data upstream and downstream simultaneously (full-duplex). The bus slots are connected to a switch that controls the data flow on the bus. A connection between a PCI Express device and a PCI Express switch is called a 'link'. Each link is composed of one or more lanes. A link composed of a single lane is called an x1 link; a link composed of two lanes is called an x2 link; etc. PCI Express supports x1, x2, x4, x8, x12, x16, and x32 link widths (lanes). The PCI Express architecture allows for a maximum bandwidth of approximately 500MB/s per lane. Therefore, the maximum potential bandwidth of this bus is 500MB/s for x1, 1,000MB/s for x2, 2,000MB/s for x4, 4,000MB/s for x8, 6,000MB/s for x12, and 8,000MB/s for x16. These values provide a significant improvement over the maximum 132MB/s bandwidth of the standard 32-bit PCI bus. The increased bandwidth support makes PCI Express ideal for the growing number of devices that require high bandwidth, such as hard drive controllers, video streaming devices and networking cards.
The usage of a switch to control the data flow in the PCI Express bus, as explained above, provides an improvement over a shared PCI bus, because each device essentially has direct access to the bus, instead of multiple components having to share the bus. This allows each device to use its full bandwidth capabilities without having to compete for the maximum bandwidth offered by a single shared bus. Adding to this the lanes of traffic that each device has access to in the PCI Express bus, PCI Express truly allows for control of much more bandwidth than previous PCI technologies. In addition, this architecture enables devices to communicate with each other directly (peer-to-peer communication).
In addition, the PCI Express bus topology allows for centralized traffic-routing and resource-management, as opposed to the shared bus topology. This enables PCI Express to support quality of service (QoS): The PCI Express switch can prioritize packets, so that real-time streaming packets (i.e., a video stream or an audio stream) can take priority over packets that are not as time critical.
Another main advantage of the PCI Express is that it is cost-efficient to manufacture when compared to PCI and AGP slots or other new I/O bus solutions such as PCI-X.
PCI Express was designed to maintain complete hardware and software compatibility with the existing PCI bus and PCI devices, despite the different architecture of these two buses.
As part of the backward compatibility with the PCI 2.2 bus, legacy PCI 2.2 devices can be plugged into a PCI Express system via a PCI Express-to-PCI bridge, which translates PCI Express packets back into standard PCI 2.2 bus signals. This bridging can occur either on the motherboard or on an external card.