Block floating-point

Block Floating Point (BFP) algorithms are a kind of algorithm used to floating point emulate while using a fixed point processor. The algorithm will assign an entire block of data an exponent, rather than single units themselves being assigned an exponent, thus making them a block, rather than a simple floating point. Block floating-point algorithm operations are done through a block using a common exponent, and can be advantageous to limit the space use in the hardware to perform the same functions as Floating-Point algorithms.