Understanding SLI and CrossFire
The term SLI was first coined by 3dfx when they released their Scan-Line Interleave technology. This technology was introduced to the consumer market in 1998 and used in the Voodoo2 line of video cards. After acquiring 3dfx, Nvidia come up with a new technology that uses the PCI-Express bus to improve graphics performance, and named it Scalable Link Interface, in short, SLI.
With the same concept as the dual core CPU, the idea behind SLI is to split the work load to smaller pieces so that different GPUs can be used to process it concurrently. In a SLI setting, two (or more) graphics cards are connected in a master-slave setting via a SLI bridge. When rendering a 3D scene, the work load is split into half and each card will be in charge of one half of the work load. When the slave card is done, it sends its output to the master card via the SLI bridge, which then combines the two results into one and sent to the output for display.
In order to use SLI on your system, you have to fulfill the following criteria:
- Own a motherboard that is based on an Nvidia chipset (the only exception is the Intel Dual Socket Extreme platform which is based on the Intel D5400XS motherboard which supports both SLI and CrossFireX technologies) and have at least two PCI-Express 16x slots.
- Two (or more) nVidia SLI graphics cards of the same GPU
SLI is rather strict on the type of GPU used. While you can set up SLI with two graphics cards from different manufacturers, the condition is: they must be using the same GPU (and preferably same clock speed and memory). In addition, not all Nvidia’s graphics cards are SLI-capable. While the latest release of graphics cards are all SLI-capable, only a selected few from the past series can run SLI. It is important to verify this before buying the graphics card.