Yes... but note that the final machine language code does use the processor opcodes that handle floats: MULSS, MOVSS http://www.felixcloutier.com/x86/MULSS.htm

Yes... but note that the final machine language code does use the processor opcodes that handle floats: MULSS, MOVSS http://www.felixcloutier.com/x86/MULSS.html http://www.felixcloutier.com/x86/MULSS.html http://www.felixcloutier.com/x86/MOVSS.html http://www.felixcloutier.com/x86/MOVSS.html So the final float multiplication is done by the hardware. What the software does, is to (a) handle the complex number and convert it to floats, and probably (b) handle the double precision using single precision x86 opcodes.