基于双精度数在计算机中是按阶码和尾码分别存储的原理,提出了一种新的计算指数函数)y=e^x的方法,它能直接计算出),的指数部分(阶码)和有效数位(尾码),并直接存储到变量),中。通过与MatLab中的库函数exp相比较,表明这种新算法计算精度更可靠,速度更快。
A new algorithm to calculate the exponential function e^x was developed to directly estimate the biased exponent and tail code based on the canister principle of the double precision number in computer. Then, the results were stored as the variant Y. Compared with the library function of exp in Maflab, it indicated the algorithm became much faster without the loss of computational accuracy.