Comparison between 2D cellular automata based pseudorandom number generators

Pseudorandom number generators (PRNGs) should satisfy two main criteria, high randomness quality and fast computation of a sequence of numbers. In this paper, a comparative study of two-dimensional Cellular Automata (CA) based PRNGs is performed to evaluate the randomness quality and the hardware constraints involved in terms of configuration parameters such as, transition rules, neighborhoods and bit extraction schemes. Experimental results show that CA-based PRNGs present good randomness quality using standard test suites, and they are well suited for parallel implementations in Field Programmable Gate Array (FPGA) technology taking advantage of the on-chip fine-grain and distributed computational resources.