PCI Express bus PCI Express (PCIe) bus came into existence at a time in which expanding slots on your computer was as crucial as the CPU’s clock speed or the amount of RAM in your system. Since then, the PCIe bus has changed from a collection of slots to plug-in extension cards to an interconnect with a high-speed topology.
The newest SSD (solid state drive) interface is an electrically PCIe 4.0 four-lane interface within the M.2 form factor. Image courtesy of w:user.snickerdo on Wikimedia Commons (CC BY-SA 3.0)
The Start of the PCIe Bus: IBM and the 5150 PC
The PCIe bus was first introduced to the IBM PC model 5150, which was released in 1981. The 5150’s predecessors, which were popular such as those of the Apple II, used open standard buses or even published their bus specifications to third-party expansion boards. The pressure from competition caused IBM to open up the 5150 bus and release its specifications.
In the early days, with IBM behind the scenes and a whole industry focused on the creation and supply of expansion cards to this IBM PC bus. The second PC model called the PC/AT, increased the bus’s bandwidth from 8 bits to 16 bits and also maintained an open architecture. Many companies utilized the bus to create clones of PCs, also known as PC compatibles. The wide usage of the bus in both the PC clone expansion boards and sectors resulted in the adoption of an industry Standard Architecture (ISA) that was used to design the bus. This was good for the consumer and PC clone makers; however, it also removed control and licensing profits from IBM.
IBM Attempts to Regain Standards Control
In the mid-80s, new processors and increased speeds were making the ISA bus obsolete. IBM launched its own exclusive microchannel busin an effort to fix the weaknesses of the ISA. IBM maintained Micro Channel proprietary to profit from licensing fees that were sold to PC-compatible manufacturers. However, the PC industry changed to Intel’s 32-bit peripheral component interconnect (PCI) bus, which is managed through PCI-SIG. PCI-SIG. (PCI-SIG).
Modern PC-compatible i80486 motherboard that was introduced in 1995 supported two formats: ISA (four slot black in the background) in addition to PCI (three white slots in the center). Photo by Duane Benson
Although PCI, like the Micro Channel PCI bus, just like that of the Micro Channel, was faster, however, it was an open standard for the entire industry. PCI was the first to introduce a technology that connected motherboard-integrated peripherals to the bus without an external card. Prior to the ISA technology, the peripherals integrated into the motherboard usually required a custom interface that was not standard circuitry. The PCI bus offered an interface to peripheral devices onboard that is electrically comparable to plugging a motherboard into a slot, which allowed an easier integration onboard and software support.
The PCI Bus (Still) Fell Short
Although PCI provided better performance than ISA, it carried several flaws from an ISA topology. As with ISA, PCI, too, the PCI bus was the shared parallel data bus. Although PCI was a significant improvement in speed and security, it required every peripheral to be able to share resources and negotiate the right to access the bus on its own.
Manufacturers of graphics accelerator cards came over these limits earlier than other manufacturers of interface cards, leading to the development of AGP, which was an accelerated graphic port (AGP). AGP was a variant of PCI that was a departure from bus sharing and offered an unimpeded connection between the AGP slot on the card with the motherboard chip.
Enter PCI Express
In 2003 the PCI-SIG also responded to these needs and introduced PCI Express, the PCI Express bus, which is still in use. PCIe was the replacement for all main PC buses, including those that use the AGP interface. PCI-SIG was the first to develop its standard in 2001 and then 2003, and PCI products started shipping in 2004.
PCIe bus PCIe bus was different from the PCI bus in two important ways. Instead of the shared bus-master topology, it utilized a point-to-point network that connects devices directly to the host controller. The system also changed from the serial data pathway to a single sequential data path.
The shared bus is part of PCI as opposed to. serial point-to-point topology in PCIe. Image taken by Matifqazi, via Wikimedia Commons (CC BY-SA 4.0)
In the previous PCI and ISA bus master topology, the only peripheral that was available at any time could connect to the bus. Each one negotiated master status when needed, waited until the master could gain control, then began to act. With Direct Memory Access (DMA), it was not possible to be accomplished in parallel. These older topologies served the requirements of slow applications from the 1980s, but they weren’t suited to gaming, high-speed networking, or the more complex graphical interfaces that are which were gaining popularity in the new millennium.
Why PCIe Was a Big Step Forward
PCIe is much more than an actual slot standard. The bus’s mainstay is its topology. PCIe connects integrated peripherals, add-in cards for mini-PCs and laptops, along with SSD storage. Mini PCIe utilizes similar topology and encryption, and specifications that are electrically compatible with standard PCIe. The increasingly popular M.2 SSD interface also uses PCIe topology.
Serial data paths used by PCIe utilize unidirectional differential pairs to increase the integrity of signals. Although these pairs must be matched in length to prevent skewing, however, the two tracks for each pair are simpler to work with than 8, 16, or 32.