Most deep learning networks support GPU computation right now, enabling the power of parallel computation capabilities.