Optimization of the Randomness Extraction Based on Toeplitz Matrix for High-Speed QRNG Post-Processing on GPU

Abstract

As one of the applications of randomness extraction, post-processing is important for the high-speed and real-time QRNG (Quantum Random Number Generator) system. This work proposes a multi-thread and streamed-processing algorithm to optimize post-processing based on Toeplitz matrix on GPU (Graphics Processing Unit) and achieves a high speed over 6.8 Gbps. Furthermore, random tests are done to evaluate the quality of the random bits generated. The results show that both the execution speed and the randomness can satisfy the requirements of the high-speed and real-time QRNG system.