A hardware gaussian noise generator and evaluation

We present a hardware Gaussian noise generator based on the Box-Muller method that provides highly accurate noise samples. The main novelty of this work is the use and precision analysis of CORDIC IPcore on a Xilinx device in computing square root and logarithmic function. The bit- widths of parameters are chosen carefully to enable rapid computation and sufficient precision. The implementation on a Xilinx Virtex-4 XC4VFX100-10 FPGA occupies 1,104 slices, 2 block RAM and 2 DSP48s. It generates two 14-bit noise samples every clock cycle and 400 million samples per second at a clock speed of 200 MHz. The performance can be improved by parallel instances. During the generation of Gaussian noise, we can also get random numbers drawn from uniform distribution, exponential distribution and Rayleigh distribution. The noise generator can be used as a key component in a hardware-based Radar echo simulator to test the sensitivity of the Radio receiver.