Energy-efficient floating-point arithmetic for digital signal processors | IEEE Conference Publication | IEEE Xplore