GPU-accelerated QPSK Transceiver with FEC over a Flat-fading Channel

Rayleigh flat-fading path in wireless-channels leads to errors, and this makes the detection task very difficult. In such cases, forward error correction (FEC) is used to provide good performance. This paper gives the testing of a QPSK-transceiver using threshold detection and FEC in the form of (8, 4) block coding-decoding. The whole system was tested by transmitting a known digital image over a flat-fading channel, and detection was performed using the threshold detection process. Very recently, the advent of programmable graphics processing units (GPUs) as excessive parallel programming system has enabled high-performance computation. NVIDIA GTX 1050 Ti GPU has been used for implementing and testing transceiver in this work. The image is transmitted over a flatfading channel along with FEC, and the results are obtained in the form of Bit Error Rate (BER) versus signal-to-noise ratio (SNR) curve. All the baseband processing is performed in the NVIDIA GPU, and some of the computation is performed in the CPU. The purpose of this paper is to show that a lot of processing time can be saved using a highly parallel computing machine, the GPU, as compared to a sequentially programming device, the CPU. The speedup is indicated in the results.