A new method of implementation of VLSI CORDIC for sine and cosine computation

In this paper, a new CORDIC method for computing sine and cosine functions with a variable scalar factor is presented. The proposed new CORDIC algorithm will reduce the number of iterations to 6 maximum, with the accuracy of 16 bit length. This is accomplished by using signed two bits, instead of one to process at a time. In order to show the feasibility and the accuracy of the proposed new CORDIC algorithm, we use C language to simulate all 2/sup 16/=65536 input angle patterns. Results have been shown that the accuracy can be up to 99% maximum and 96% minimum.