CNC interpolation is essentially a numerical approximation in mathematics. The Bernstein polynomials have played an important role in the area of approximation ever since it was created. Hence the study on their rapid computations is valuable in CNC interpolation. Classical studies on rapid computations always focus on improving the computational efficiency by software algorithms and seldom associate the problem with the working process of the computer's processor. This paper introduces a different thought. The paper presents a method that can rapidly compute the cubic Bernstein polynomials in accordance with the working traits of computers' processors. Consequently, the computational routine can be implemented both by software and by hardware, which greatly enhances the computational efficiency. The principles and process of the method are presented together with its application in shape optimization.