An efficient design method of four parallel channels in-order FFT with single floating-point butterfly is proposed, to reduce the resource consumption and improve the real-time calculation ability. The radix-4 FFT is deduced to calculate the access address for four channel data parallel. The hardware architecture of the proposed FFT is presented, and the single precision floating-point adder and multiplier are also depicted. The proposed architecture of a four channels 1024 points radix-4 FFT with single butterfly is implemented in FPGA, and the performance is compared with previous literatures and some EDA corporations IP cores, which shows the correctness and effectiveness of the proposed method.
A reconfigurable radix-r FFT structure design is proposed, to reduce the memory and time consumption caused by zero-paddling in traditional radix-2/4 methods. The radix-r FFT computing flow is divided into three iterative steps: rotate factors computation; memory access schedule; butterfly coefficients matrix multiplication. The hardware structure is depicted, in which the memory accessing schedule is implemented by a finite state machine, the rotate factors and butterfly coefficients are calculated by the Taylor series expansion. The time consumption of the proposed structure is analyzed, and with the arbitrary data points, an optimized radix can be used to minimize the computing time. Compared with traditional methods, the proposed design has the advantage of generality, less time consumption and hardware costs.
A dual-butterfly parallel access constant geometry pipeline radix-2 FNT (Fermat Number Transform) is proposed to enhance the computing performance of FNT. By the extending the conventional constant geometry FNT, two radix-2 butterflies could be calculated simultaneously in each stage, and the address generating method for parallel access without conflicts is deduced to make the dual-butterfly’s four operators fetched and stored at the same time. Compared with other single data stream FNT, the efficiency is enhanced by 3 times. Compared with the traditional convolution and the convolution based on conventional FNT, the convolution based on the proposed algorithm has the advantage in computing efficiency, which also indicates the efficiency of the proposed algorithm.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.