MIMO系统固定复杂度球形解码器研究

Research of Fixed-Complexity Sphere Decoder for MIMO Systems on CUDA

图形处理单元(GPU)可作为低成本并行可编程协处理器,能提供高的计算吞吐量,非常适于大规模系统设计。为充分利用GPU的并行处理能力,采用专为GPU通用计算所设计的CUDA(Compute Unified Device Architecture)架构,针对平坦衰落信道,设计了用于MIMO无线通信系统的固定复杂度球形解码器,以加速MIMO无线通信系统的数据处理。根据GPU的架构与存储特点,对实现进行了优化,减小了数据存取延迟和访问冲突。实验结果表明,球形解码速度可提高近10倍.

Graphic Processing Units（GPUs） is a low-cost parallel programmable co-processor that can deliver extremely high computation throughput and is well suited for large-scale system design.In this paper,sphere decoding algorithm on CUDA which utilize the parallel processing power of GPU to accelerate the data processing of MIMO systems is investigated,flat fading channel is considered and an optimized fixed-complexity sphere decoder on GPU is implemented,and optimization is performed according to GPU memory utilities to reduce data transfer delay.Performance results show that the execution of the sphere decoder can be accelerated by factors of over 10.