SMUL-FFT: A Streaming Multiplierless Fast Fourier Transform

Beamspace processing is an emerging paradigm to reduce hardware complexity in all-digital millimeter-wave (mmWave) massive multiple-input multiple-output (MIMO) basestations. This approach exploits sparsity of mmWave channels but requires spatial discrete Fourier transforms (DFTs) across the antenna array, which must be performed at baseband sampling rate. To mitigate resulting DFT implementation bottleneck, we propose a fully-unrolled Streaming MUltiplierLess (SMUL) fast Transform (FFT) engine that performs one transform per clock cycle. The proposed SMUL-FFT architecture avoids multipliers by restricting twiddle factors sum-of-powers-of-two, substantial power and area savings. Compared state-of-the-art FFTs, our ASIC designs 65nm CMOS demonstrate more than 45% 17% improvements energy-efficiency area-efficiency, respectively, without noticeably increasing error-rate MIMO systems.