LDPC Decoding via Processing-in-Memory for Mobile and Wireless Communications
Mojtaba Mahdavi · 2025
Low-density parity-check (LDPC) codes are widely used in modern communication systems, including 5G networks, due to their near-capacity error correction performance. However, conventional LDPC decoding implementations suffer from high memory bandwidth demands and computational complexity, particularly in high-throughput applications. Processing-in-memory (PIM) has emerged as a promising computational paradigm to overcome the memory bottleneck by integrating computation within memory, thereby reducing data movement overhead. In this paper, we explore the realization of LDPC decoding using PIM architectures, where LDPC decoding computations are mapped onto arrays of memristor devices. Analytical results demonstrate that our proposed PIM-based LDPC decoder significantly improves throughput and latency compared to conventional architectures, highlighting the potential of PIM for next-generation error correction systems.