Floating-point operations are critical for many applications where heterogenous computing is gaining traction. Floating-point implementation can be power hungry, however. As such, this thesis presents the architecture and ...