Does anyone know what is the design of a "saturating adder." I was reading a digital design book and I've come across some specific examples online but I was wondering how an n-bit saturating adder works and why is it useful.

 

A saturating adder is used in digital signal processing applications. It differs from an ordinary 2s complement adder by not allowing overflow or underflow in fixed point operations. Instead it satuarates like an analog adder whose inputs cause the output to go to the supply rails.

0x7FF0 + 0x0020 = 0x8010
In a 2s complement adder this would cause what should be a positive result to return a negative number
In a saturating adder
0x7FF0 + 0x0020 = 0x7FFF
brcause 0x7FFF is the maximum positive or 'saturated' value

Does that clear things up?