Data-parallel implementation of reconfigurable digital predistortion on a mobile GPU

3GPP LTE-A offers new technologies such as non-contiguous carrier allocation for improving radio spectrum utilization. However, implementation of these technologies is challenging because of intermodulation distortion caused by non- linearity of components. Digital Predistortion (DPD) offers a way for compensating for these nonlinearities by modifying the digital baseband signal. As most consumer-oriented mobile devices are equipped with powerful Graphics Processing Units (GPUs), it has become possible to implement DPD functionality to such devices with no additional hardware cost. In this paper, we propose data- parallel, reconfigurable predistortion and measure its performance on mobile GPUs: Qualcomm Adreno 330 and ARM Mali T628.