Comparative analysis of radix-2, radix-4, radix-8 CORDIC processors

Abstract

Digital Signal Processing is widely used in number of data streaming applications. The analytical treatment of all DSP algorithms involve basic arithmetic operations like multiplication and division besides the transcendental functions viz. trigonometric, inverse trigonometric, logarithmic, exponential etc. Generally used software solutions for digital implementation of these products are table look up or Taylor series expansion, but these require large memory and calculation become tedious. Coordinate rotational Digital Computer is one such technique which uses just shift and add arithmetic to implement different functions. By varying few simple parameters, it can be used to efficiently implement trigonometric, logarithmic, exponential function etc. Since it uses only shift and add arithmetic, VLSI implementation of such an algorithm is achievable. Radix-2 CORDIC Processor requires n iteration for n bit input data, so the latency is large. So to reduce the latency higher radix i.e. radix-4 and radix-8 can be employed, which reduces the number of iterations from n/2 to n/4. It is observed from the simulation and synthesis result that radix-8 is more energy efficient and gives high throughput compared to that of radix-2 and radix-4 CORDIC architectures with extra area overhead.