Summary
International Symposium on Nonlinear Theory and Its Applications
2022
Session Number:A2L-B
Session:
Number:A2L-B-02
Implementation of Secure and Fast Pseudo-Random-Number Generator on GPU
Hitoaki Yoshida , Takeshi Murakami,
pp.5-8
Publication Date:12/12/2022
Online ISSN:2188-5079
Summary:
A novel iteration method with the modified activation function has been proposed to implement the secure and fast pseudo-random number generator (PRNG). The modified activation function has accelerated the pseudo-random-number (PRN) generation rate to 56.8 Tb/s by using NVIDIA GPU (A100). The PRNG consists of 1.67× 107 Chaotic and Random Neural Networks (CRNNs), and the whole period of the PRN series is estimated at P～10^{100000000} based on the L.C.M. of the 10000 PRN series. The result is expected to apply the information security of the Controller Area Network (CAN) and that of the Autonomous GPU Accelerated systems, especially intelligent vehicle systems.