Canonical-signed-digit (CSD) coefficient representations of floating point numbers lead to efficient add/subtract networks for hardwired DSP multiplies of two?s complement signals.

example csdigit(23) returns +0-00-.
Where +0-00-. is a representation of +1 in 2^5, -1 in 2^3 and -1 in 2^0 positions.
i.e. 23 = 32 - 8 - 1

example [a,p,n]=csdigit(23.5,6,2) returns
a = +0-000.-0
p=32
n=8.5
23.5 = 32 - 8 - 0.5