Application of user-defined floating-point formats | IEEE Conference Publication | IEEE Xplore