High-Throughput Pseudo-Random Number Generators Over Discrete Chaos

In this brief we propose two pseudo-random number generators (PRNGs) based on the discrete Arnold cat map (DACM) over the integer ring $\mathbb {Z}_{2^{m}}$ . The first PRNG employs or-exclusive (XOR) and bit permutations, discarding multiplication operations. The second PRNG proposes a methodology to double the number of bits extracted from each chaotic sample, increasing the bit generation rate. The statistical properties of the PRNGs are analyzed with the statistical test suites NIST and TestU01. The proposed PRNGs are implemented in the field-programmable gate array (FPGA) Xilinx Zynq-7000 and their hardware complexity is analyzed. We show that the maximum throughput obtained by one of the proposed PRNGs is 19 Gbps, which outperforms recent approaches.