SLI and CrossFire, a necessary analysis
I remember perfectly that it was 3DFX which marked the starting point of the multi-GPU configurations, with the release of the Voodoo 2 in the general consumer market, as graphical solution that can be considered the origin of what would later become SLI and CrossFire, made by NVIDIA and AMD respectively.
Originally SLI (Scan Line Interleave) mode of the Voodoo 2 acted in a simple way: two cards running in parallel resorting to a division of the processing load simply but effectively, dividing the screen in two and distributing the lines 50-50.
As is the case with the current solutions at that time, this meant doubling the power of rendering, but the texture memory is not added for a very simple reason which is that each graphics card needs to double the scene data it’s working on.
This is a very important limitation that we currently still have, just like the level of support on the software side and many others that we’ll review in this special, where we will give a series of basic guidelines that will allow you to better understand how SLI and CrossFire modes work and assess whether or not they are really worth it.
A preview
Before we jump in, we will explain a little better what these technologies are and how they work, as that will provide us the necessary foundation on which we can have a better understanding of everything that we will explain below. Please note that this article is intended to be useful even for the most amateurish, so it is likely that you find plenty of information you may already know.