Accelerate GMSK demodulation using GPU

Gaussian Filtered Minimum Shift Keying(GMSK) is used extensively in radio communication field nowadays; as frequency of radio signals increases, traditional technology such as DSP and FPGA can´t satisfy the real-time requirement due to its complexity and compute-intensive. The development of modern GPU makes this possible. In this paper, a Gaussian Filtered Minimum Shift Keying (GMSK) demodulation was implemented using CUDA on GPU. Since the demodulation process was largely executed on the GPU, the method can take full advantage of massive computational power of modern graphics hardware. The experiment results show that our GPU based GMSK implementation can gain nearly 20× speedups compared with the serial algorithm.