The floating point segment of a processor, is that which deals with numbers that aren't straight integers - in other words, numbers that have something other than zero after the decimal point. All modern processors come with an onboard floating point unit, usually referred to as the FPU, and this deals with the complex calculations needed by specific tasks like generating 3D graphics or working with complex mathematical data. The basic point is that a processor with an efficient FPU will perform better at intensive number-crunching jobs than one without. Generally, Intel's processors used to have the most efficient floating-point performance, although that changed, as AMD's Athlon 64 FPU is far better in terms of performance.