高速量子随机数产生中的实时并行后处理

Real-Time and Parallel Post-Processing for High-Speed Quantum Random Number Generation

针对目前串行量子随机数发生器方案中,量子模式提取的信号带宽与数据处理速率之间的权衡在实时产生速率方面造成了瓶颈,文章利用现场可编程门阵列(FPGA)对多个量子边带模式的并行托普利茨(Toeplitz)哈希实时后处理,研究分析了不同规模Toeplitz矩阵的逻辑资源占用,设计了两层并行流水线算法并进行了仿真验证,充分利用FPGA的并行处理优势和硬件资源,最终实现了实时速率8.24 Gbit/s的量子随机数安全后处理。仿真结果表明,量子随机数发生器的熵源具有较好的随机性,经后处理后的随机数自相关系数保持在10-4量级且能通过Diehard测试。通过与集成电路匹配并行提取多个量子边带模式,实现了一种可商业化的高鲁棒性、高速实时量子随机数发生器。

At present,the tradeoff between electronic bandwidth in quantum mode extraction and data processing rate create a bottleneck in real-time generation ratio in such serial-type Quantum Random Number Generator(QRNG)scheme.To solve the problem,we investigate parallel Toeplitz-hashing post-processing of multiple quantum sideband modes within one Field Programmable Gate Array(FPGA).Logic source occupation for different Toeplitz matrices is analyzed and two-layer parallel pipeline algorithm is delicately designed to fully exploit the advantage of parallel algorithm and hardware source of the FPGA.A simulation is conducted to verify the design.This work finally achieves a real-time random number generation rate about 8.24 Gbit/s.The experimental results show that the entropy source of the quantum random number generator has good randomness.After the post-processing,the coefficient of the autocorrelation function of the random numbers remain in the order of 10-4 and the random numbers can pass the Diehard test.Matching up with integrated circuit for parallel extraction of multiple quantum sideband modes,our demonstration shows a robust,extensible,high-speed and real-time QRNG,which has the potential for commercialization.