基于X64架构的泛探雷达并行处理系统设计

Design of Parallel Processing System for Pan-detector Radar Based on X64 Architecture

针对“低慢小”目标探测的泛探雷达仿真系统存在运行时间长运算效率低的问题,本文提出了一种双层并行处理技术,每一层都在存储上使用了双缓冲模型,在时序上使用“生产者-消费者”模型。该技术以增加一倍内存开销为代价换取了时间上“生产”和“消费”的同步执行,从而降低系统运行时间。经过仿真验证以及对比,该技术相比传统串行处理技术,系统运行时间降低了73%。

Addressing the problem of long running time and low operation efficiency of the simulation system of the Pan-detector detection radar based on“low slow and small”target detection.A double-layer parallel processing technique was proposed,each layer used a dual buffering model on storage,and a“producer-consumer”model on timing.This technology traded the synchronous execution of“production”and“consumption”in time at the cost of doubling the memory cost,thus reducing the system running time.After simulation and comparison,the system running time of this technology was reduced by 73%compared with the traditional serial processing technology.