Floating-point computation of functions with maximum accuracy | IEEE Conference Publication | IEEE Xplore