基于GPU并行计算技术的大约束长度卷积码的恶码检验

卷积码生成多项式对通信系统中信道编译码的性能有决定性影响,通常我们利用计算机搜索大约束长度的卷积码生成多项式,在搜索得到的生成多项式中有相当数量是恶性的需剔除,因此必须对搜索到的生成多项式进行检验.判定一个编码器的生成多项式是否是恶性的方法可以是数学推导或Matlab计算.但只适合约束长度较短的卷积码,而当约束长度较大时,运用现有的计算手段很难对编码器生成多项式进行检验.本文提出了一种新的基于GPU的高性能并行运算技术验证卷积码生成多项式的方法,GPU在进行大约束长度卷积码的恶码检验上有明显的速度优势.

Inspection of the Long Constraint Length Convolution Code about the Evil Code Based on GPU Parallel Computing Technology

In order to ensure the reliability of a communication system and to decrease the noise and interference, channel coding should be introduced in ＂to＂ communication system. When using the computer to search a long constraint length of the convolution code generator polynomial, the genera- tor polynomial must be tested to remove the malignant . We use mathematical and Matlab to deter- mine whether the encoder generator polynomial is vicious. However, these methods are only suitable for the short constraint length convolutional code, when the constraint length is longer, it is difficult to test the generator polynomial of the encoder with existing computing means. This paper puts for- ward a new method of verifying the convolution code generation polynomial which is based on the high performance of GPU＂ , the experimental simulation results show that this technology can greatly increase the operation efficiency and has high practical value. GPU has obvious advantages in the e- vil code inspection of the large constraint length convolutional code.