基于GPU的高性能并行计算技术

High Performance Parallel Computing Technology Based On GPU

摘要:

针对NVIDIA CUDA(Compute Unified Device Architecture)架构的第三代GPU高性能计算技术开展了研究,利用具有448个处理核心的NVIDIA GPU GTX470实现了脉冲压缩雷达的基本数据处理算法,包括脉冲压缩算法与相参积累算法。根据GPU的并行处理架构,将脉冲压缩、相参积累完成了并行算法优化设计,有效的将算法映射到GPU GTX470的448个处理核心中,完成了脉冲压缩雷达基本处理算法的GPU并行处理实现,并针对处理结果效果与实时性进行了评估。

Abstract:

The research focus on high-performance computing technology based on the third generation GPU of NVIDIA CUDA (Compute Unified Device Architecture).Pulse compression radar data processing algorithms including of pulse compression and coherent accumulation have been realized by using GPU GTX470 which has 448 processing core. According to GPU parallel processing architecture, algorithms of pulse compression and coherent accumulation have been optimized and mapped to the 448 processing core of GPU TTX470 effectively, the pulse compression radar processing algorithm has been implemented based on GPU, the result and processing real-time performance have been evaluated.