Optimization Of Polarization Encoding And Decoding Algorithm Based On GPP Platform

Since the birth of modern communication technology,how to achieve more reliable communication with lower cost(signal power,signal bandwidth,transmission delay,etc.)has been the core key in the field of information technology,and the most core problem in wireless communication system is channel coding technology.Polar Code,which is based on the theory of channel polarization,is the first channel coding method that can be proved strictly to reach shannon channel capacity in the field of communication,and its error correction performance is also better than Turbo Code,LDPC Code and other decoding methods.On the other hand,universities and research institutes usually choose to use software to build a simulation experiment platform to assist scientific research when conducting research on communication systems.Therefore,this paper also chooses a general processor platform for research.In this paper,a series of concepts of polarization codes derived from the theory of channel polarization are studied,and the classical coding algorithms and implementation principles of polarization codes are analyzed.Then it introduces the features and advantages of the general processor platform,the main optimization methods and performance analysis tools.Then from size polarization encoding decoding algorithm simulation system,using a variety of parallel optimization method to complete the experiment coding and optimization work,and then use the performance analyzer for auxiliary,locate hotspots and implement targeted optimization,so as to achieve the validity and feasibility of this thesis about the optimization scheme.The main work and innovation points of this paper include the following:First,this paper proposes a fast and simple continuous deletion maximum likelihood decoding algorithm(FSSC-ML),and carries out parallel optimization using AVX-512,and the simulation verifies the effect before and after optimization.Secondly,by reasonably using Click Plus,OpenMP and TBB three parallel programming models to add parallel structure to the code,to parallel the loop,sequence,recursive function and other code patterns,to achieve programming parallelization.Finally,by making full use of the characteristics of the general processor platform,this paper analyzes the characteristics of different algorithm modules and selects table lookup method,high-performance function library and single-instruction multi-data flow instruction set to increase the parallelism of the system.By combining various parallel processing acceleration technologies,the simulation experiment of polar-code coding and decoding algorithm optimization can meet the performance requirements of real-time processing on ordinary laptop.The experimental results show that the optimization scheme proposed in this paper can effectively reduce the system processing delay,well meet the real-time signal processing delay requirements of wireless communication system,and achieve good results,which is of great significance for future scientific research.