FeMBMS eNB / UE End-to-End chain; NR: inproved modulation speed, CUDA GPU ldpc latency