An apparatus (100) for computing the absolute value of a complex number includes separate squaring units (110, 115) for the real and imaginary parts. A square root unit (130) extracts the square root of the sum (120) of these squares, which is absolute value of the complex number. Each squaring unit includes one unsigned multipliers for respective least significant and two signed multipliers for respective most significant bits and a cross term. The products are aligned by shifting and summed. The square root unit employs identical processing elements, each considering two bits of the input and forming one root bit and a remainder. Each processing element compares two intermediate test variables, and selects a “1” or “0” for the root bit and the next remainder based upon this comparison. A chain of processing elements enables computation of the root to the desired precision. Alternatively, the same processing elements may be used in a recirculating manner.