Introduction

Intel has certainly led the field in many of the technologies we now take for granted. PCI, USB, AGP and SDRAM are all projects that Intel has pushed into the market. Some have been greeted with great enthusiasm like the AGP, PCI and USB buses, some were a groaning point for manufacturers, such as re-designing motherboards to use SDRAM rather than SIMM modules, and some were totally off the charts bad - such as the now extinct RDRAM modules better known as RAMBUS.

PCI Express is the latest to come from the Intel labs. A joint venture with many companies to move to a more unified bus system, PCI Express has already been getting a lot of hype, especially with its capabilities to totally replace AGP by allocating over 8GB/s of bandwidth, rather than the 2.1GB/s which AGP 8x supplies.

While this seems good in theory, this leads to many new problems, such as new graphics slots, re-designed boards, and of course, the GPU manufacturers to support the new interface protocol.

When AGP came in many years ago, support was easy as AGP was simply a ramped up PCI bus. Over the years it has added new features, however, in its initial introduction, AGP offered only 266MB/s data transfer running a dedicated PCI subsystem at 66MHz rather than 33MHz. PCI Express is a totally different protocol. Its nature is serial, rather than parallel. It uses a totally different signalling process, which means that AGP based graphics cards aren't able to communicate with PCI Express natively, like PCI graphics chipsets were able to do on AGP slots.

Today we are taking a look at nVidia's first PCI Express based graphics cards to hit the market in the mid to mainstream range cards, based on existing technologies. Please join us as we take a look at the PCX5900 from Gigabyte and PCX5750 from Albatron.

Find the lowest price on PCI Express Graphics Cards!