Floating point processor
From Computer History Wiki
A floating point processor was, until recently, an option which allowed a computer to handle floating point numbers in hardware. In early machines, it was a complete extra set of boards; later on, with the rise of microprocessors, it became an optional chip.