| 引用本文: | 金莉,孔文青,宋万杰. 基于GPU的信号产生及脉冲压缩实现[J]. 雷达科学与技术, 2017, 15(5): 505-508.[点击复制] |
| JIN Li, KONG Wenqing, SONG Wanjie. Signal Generation and Pulse Compression Realization Based on GPU[J]. Radar Science and Technology, 2017, 15(5): 505-508.[点击复制] |
|
| 摘要: |
| 文中采用了一种基于CPU+GPU异构并行架构体系的信号处理方案。按照雷达信号处理流程，通用处理计算机利用CPU串行代码完成核函数启动前数据准备和设备初始化工作，并控制信号处理的任务调度和负载分配，然后将数据通过PCI E总线传输至显存，利用GPU特有的单指令多线程方式，并行实现线性调频信号产生以及线性调频信号频域脉冲压缩算法，并与CPU进行比较。实验结果表明，利用计算统一设备架构技术实现的线性调频信号产生以及脉冲压缩算法取得了比CPU更高的运算效率。 |
| 关键词: 计算统一设备架构 图形处理机(GPU) 线性调频信号 脉冲压缩 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金(No.61571349) |
|
| Signal Generation and Pulse Compression Realization Based on GPU |
|
JIN Li, KONG Wenqing, SONG Wanjie
|
|
(National Key Laboratory of Radar Signal Processing, Xidian University, Xi’an 710071, China)
|
| Abstract: |
| A signal processing scheme based on CPU+GPU heterogeneous parallel architecture is proposed in this paper.According to radar processing flow, the general computer uses CPU to complete data preparation and device initialization before kernel function starting, and control the task scheduling and load sharing of signal processing, then, the data will be transmitted through the PCI-E bus to the memory of GPU.The specific way of SIMT of GPU is used to achieve thread parallelization, and finally finish the LFM signal generation and pulse compression algorithm.A comparison with the algorithm realized on CPU is made.The experimental results show that the algorithm achieved by CUDA has higher efficiency than CPU. |
| Key words: compute unified device architecture(CUDA) graphical processing unit(GPU) LFM signal pulse compression |
