DecodeX: Exploring and Benchmarking of LDPC Decoding Across CPU, GPU, and ASIC Platforms
Emerging virtualized radio access networks (vRANs) demand flexible and efficient baseband processing across heterogeneous compute substrates, where low-density parity-check (LDPC) decoding in the Physical layer remains one of the most performance- and energy-critical workloads. In this paper, we present DecodeX, a unified benchmarking framework for evaluating LDPC decoding implementations across different hardware platforms. DecodeX integrates a comprehensive suite of LDPC decoder implementations, including kernels, APIs, and test vectors for CPUs (FlexRAN), GPUs (Aerial and Sionna-RK), and ASICs (ACC100), and can be readily extended to additional architectures and configurations. Using DecodeX, we systematically characterize how different platforms orchestrate computation–from threading and memory management to data movement and accelerator offload–and quantify the resulting decoding latency under varying Physical layer parameters. Our observations reveal distinct trade-offs in parallel efficiency and offload overhead, showing that accelerator gains strongly depend on data-movement and workload granularity. Building on these insights, we discuss how cross-platform benchmarking can inform adaptive scheduling and co-design for future heterogeneous vRANs, enabling scalable and energy-efficient baseband processing for NextG wireless systems.
