Abstract
Memory-based architectures have received great attention for single-chip implementation of the fast Fourier transform (FFT). Basically, they can be roughly categorized as single-memory design, dual-memory design, and buffer-memory design. Among them, the buffer-memory design can balance the trade-off between memory size and control circuit complexity. In this paper, we present a design methodology of buffer-memory architectures for the radix-2 decimation-in-frequency FFT algorithm that can effectively reduce the needed memory. As compared to previous related works, the designs derived from the proposed methodology can reach the same throughput performance with a smaller memory size. These designs are rather attractive for long-length FFT applications, such as very-high-rate digital subscriber lines and digital video broadcasting.