Programmable Dataflow Accelerators: A 5G OFDM Modulation/Demodulation Case Study
Published in ICASSP 2020 - IEEE International Conference on Acoustics, Speech, and Signal Processing, 2020
This paper investigates the design and implementation of programmable dataflow accelerators tailored for 5G Orthogonal Frequency Division Multiplexing (OFDM) modulation and demodulation. The authors propose a novel architecture that leverages dataflow programming models to achieve high throughput and energy efficiency, addressing the challenges posed by the increasing demand for data rates in 5G wireless communication systems.
Key contributions include:
Dataflow Programming Model: The introduction of a dataflow-based programming model that allows for efficient mapping of OFDM modulation and demodulation tasks onto hardware accelerators.
Performance Evaluation: Comprehensive performance evaluation demonstrating significant improvements in throughput and energy efficiency compared to traditional processing methods.
Scalability: Analysis of the scalability of the proposed architecture, highlighting its suitability for future 5G and beyond communication systems.
Recommended citation: Wu, Y., Wang, P., & McAllister, J. (2020). Programmable Dataflow Accelerators: A 5G OFDM Modulation/Demodulation Case Study. In *ICASSP 2020 - IEEE International Conference on Acoustics, Speech, and Signal Processing* (pp. 1728–1732). IEEE. https://doi.org/10.1109/ICASSP40776.2020.9053529
Download Paper | Download Slides | Download Bibtex