水下无线光信道并行仿真方法研究  

Research on Parallel Simulation Method of Underwater Wireless Optical Channel

作　　者：张建磊 寇琳琳 王洁 刘雪辰 杨祎 贺锋涛 Zhang Jianlei;Kou Linlin;Wang Jie;Liu Xuechen;Yang Yi;He Fengtao(College of Electronic Engineering,Xi’an University of Post&Telecommunications,Xi’an 710121,Shaanxi,China;Naval Equipment Department,Xi’an 710077,Shaanxi,China)

机构地区：[1]西安邮电大学电子工程学院,陕西西安710121 [2]海军装备部,陕西西安710077

出　　处：《激光与光电子学进展》2022年第19期105-114,共10页Laser & Optoelectronics Progress

基　　金：装备预研教育部联合基金(8091B032130);国家自然科学基金(61805199);陕西省技术创新引导专项基金(2020TG001);水下信息与控制重点实验室基金。

摘　　要：针对基于蒙特卡洛的水下无线光信道仿真计算量大、计算效率低的问题,提出了利用Open MP(open multiprocessing)和CUDA(compute unified device architecture)的水下无线光信道并行仿真方法。通过将计算密集型部分移植到各线程并行计算的方式提高仿真计算效率。在此基础上引入3种优化方案,通过剔除无效光子和限制高散射事件的方式,加速数据合并,减少主存显存的数据交换量,进一步提高仿真效率。对比分析了在不同水质、不同计算环境以及不同光子数和距离等条件下的加速效果。结果表明,相比于传统串行仿真,图像处理器(GPU)水下光散射并行计算方法的加速最高可达300倍;中央处理器(CPU)水下光散射计算方法的加速最高可达90倍。A parallel simulation method for underwater wireless optical channel simulation is proposed based on open multiprocessing(Open MP)and compute unified device architecture(CUDA)to address the computational complexity and low computational efficiency of the Monte Carlobased model.The simulation calculation efficiency is improved by transplanting computationally intensive parts to threaded parallel computing.Three optimization schemes are introduced to accelerate merging by eliminating invalid photons,limiting highscattering events,and reducing the amount of data exchange between main and video memories,further improving the simulation efficiency.The acceleration effects in various water types,computing environments,and photon numbers and distances are evaluated.The parallel calculation efficiency of graphics processing units has an acceleration effect of up to 300 times,whereas that of central processing units has an acceleration effect of 90 times compared with traditional serial simulation.

关 键 词：海洋光学 并行架构 水下无线光通信 蒙特卡洛仿真 

分 类 号：TN911.74[电子电信—通信与信息系统]