Pakize ERGÜL, H. Fatih UĞURDAĞ, Doğancan DAVUTOĞLU

HC-FFT: highly configurable and efficient FFT implementation on FPGA

is one of the basic building blocks in many applications such as sensors, radars, communications. For some applications, e.g., real-time spectral monitoring and analysis, FFT needs to be ”run-time configurable” so that the system is real-time. When examining the previous work on configurable real-time (FPGA-based) FFT implementations, we see that the degree of configurability is less than what is desired. In this paper, a new FFT architecture is proposed, which has a high degree of run-time configurability and yet does not compromise area or throughput. The configurable parameters of this design are the number of FFT points (up to 64K), forward versus inverse mode, output order (natural or bit-reversed), and the number of streams (up to 4). The proposed FFT architecture (HC-FFT) is designed using a parallel and pipelined radix-2 multipath delay commutator (MDC) FFT structure. HC-FFT was implemented on a Xilinx Kintex Ultrascale FPGA and was verified against the Xilinx FFT IP. Besides its high degree of run-time configurability, HC-FFT is quite efficient and offers a very high throughput of 87 Gbps with a quite reasonable area.

PDF

___

[1] Garrido M, Grajal J, Sanchez MA, Gustafsson O. Pipelined radix-2 k feedforward FFT architectures. IEEE Transactions on Very Large Scale Integration Systems 2013; 21 (1): 23-32. doi: 10.1109/TVLSI.2011.2178275
[2] Chen S, Huang S, Garrido M, Jou S. Continuous-flow parallel bit-reversal circuit for MDF and MDC FFT architectures. IEEE Transactions on Circuits and Systems I: Regular Papers 2014; 61 (10): 2869-2877. doi: 10.1109/TCSI.2014.2327271
[3] Yoshizawa S, Orikasa A, Miyanaga Y. An area and power efficient pipeline FFT processor for 8×8 MIMO-OFDM systems. In: IEEE International Symposium of Circuits and Systems; Rio de Janeiro, Brazil; 2011. pp. 2705-2708.
[4] Boopal PP, Garrido M, Gustafsson O. A reconfigurable FFT architecture for variable-length and multi-streaming OFDM standards. In: IEEE International Symposium on Circuits and Systems; Beijing, China; 2013. pp. 2066-2070.
[5] Li S, Xu H, Fan W, Chen Y, Zeng X. A 128/256-point pipeline FFT/IFFT processor for MIMO OFDM system IEEE 802.16e. In: IEEE International Symposium on Circuits and Systems; Paris, France; 2010. pp. 1488-1491.
[6] Cooley JW, Tukey J. An Algorithm for the machine calculation of complex Fourier series. Mathematics of Computation 1965; 19 (90): 297-301. doi: 10.2307/2003354
[7] Kumar GP, Krishna BT, Pushpa K. Optimized pipelined fast Fourier transform using split and merge parallel processing units for OFDM. Wireless Personal Communications 2020; 117 (4). doi: 10.1007/s11277-020-07471-3
[8] O’Brien J, Mather J, Holland B. A 200 MIPS single-chip 1 k FFT processor. In: IEEE International Solid-State Circuits Conference; New York, NY, USA; 1989. pp. 166-167.
[9] Sunada G, Jin J, Berzins M, Chen T. COBRA: an 1.2 million transistor expandable column FFT chip. In: IEEE International Conference on Computer Design: VLSI in Computers and Processors; Cambridge, MA, USA; 1994. pp. 546-550.
[10] Baas BM. A low-power, high-performance, 1024-point FFT processor. IEEE Journal of Solid-State Circuits 1999; 34 (3): 380-387. doi: 10.1109/4.748190
[11] He S, Torkelson M. Design and implementation of a 1024-point pipeline FFT processor. In: IEEE Custom Integrated Circuits Conference; Santa Clara, CA, USA; 1998. pp. 131-134.
[12] Langemeyer S, Pirsch P, Blume H. A FPGA architecture for real-time processing of variable-length FFTS. In: IEEE International Conference on Acoustics, Speech and Signal Processing; Prague, Czechia; 2011. pp. 1705-1708.
[13] Xie Y, Feng Y, Yang C, Xie Y, Chen H. Design of a large point FFT processor with configurable transform length. In: IET International Radar Conference; Hangzhou, China; 2015. pp. 1-5.
[14] Gautam V, Ray KC, Haddow P. Hardware efficient design of Variable Length FFT Processor. In: IEEE International Symposium on Design and Diagnostics of Electronic Circuits and Systems; Cottbus, Germany; 2011. pp. 309-312.
[15] Karachalios A, Nakos K, Reisis D, Alnuweiri H. A new FFT architecture for 4 × 4 MIMO-OFDMA systems with variable symbol lengths. In: International Conference on Innovations in Information Technology; Al Ain, United Arab Emirates; 2009. pp. 80-84.
[16] He C, Qiang W, Zhenbin G, Hongxing W. A pipelined memory-efficient architecture for ultra-long variable-size FFT processors. In: International Conference on Computer Science and Information Technology; Singapore; 2008. pp. 357-361.
[17] Manish B, Sangeeta N. Fast performance pipeline re-configurable FFT processor based on radix-2 2 for variable length N. International Journal of Electrical and Electronic Engineering and Telecommunications 2019; 8 (3): 163- 170. doi: 10.18178/ijeetc.8.3.163-170
[18] Netto R, Güntzel JL. A high throughput configurable FFT processor for WLAN and WiMax protocols. In: VIII Southern Conference on Programmable Logic; Bento Goncalves, Brazil; 2012. pp. 1-5.
[19] He J, Ma L, Xu X. A configurable FFT processor. In: IET International Conference on Wireless, Mobile and Multimedia Networks; Beijing, China; 2010. pp. 246-249.
[20] Yang C, Xie Y, Chen L, Chen H, Deng Y. Design of a configurable fixed-point FFT processor. In: IET International Radar Conference; Hangzhou, China; 2015. pp. 1-4.
[21] Cho I, Shen C, Tachwali Y, Hsu C, Bhattacharyya SS. Configurable, resource-optimized FFT architecture for OFDM communication. In: IEEE International Conference on Acoustics, Speech and Signal Processing; Vancouver, Canada; 2013. pp. 2746-2750.
[22] Eddington C, Ray B. Using the parallel FFT for multigigahertz FPGA signal processing. Xcell Journal 2013; 82: 51-55.
[23] Iglesias V, Grajal J, Sánchez MA, López-Vallejo M. Implementation of a real-time spectrum analyzer on FPGA platforms. IEEE Transactions on Instrumentation and Measurement 2015; 64 (2): 338-355. doi: 10.1109/TIM.2014.2344411
[24] Verma R. FPGA implementation of fast Fourier transform (FFT) based finite impulse response (FIR) filter using VHDL. International Journal of Engineering and Computer Science 2014; 3 (3).
[25] Wang S, Inkol R, Rajan S, Patenaude F. FFT filter-bank-based wideband detection: coherent vs. non-coherent integration. In: IEEE Instrumentation and Measurement Technology Conference; Singapore; 2009. pp. 1327-1331.
[26] Xilinx Corp. Fast Fourier Transform v9.1 LogiCORE IP Product Guide PG109. San Jose, CA, USA: Xilinx Inc., 2020.
[27] Bruno JS, Almenar V, Valls J. FPGA Implementation of a 10 GS/s variable-length FFT for OFDM-based optical communication systems. Microprocessors and Microsystems 2019; 64: 195-204. doi: 10.3390/electronics7070116
[28] Nash JG. Distributed-memory-based FFT architecture and FPGA implementations. Electronics 2018; 7: 116. doi: 10.1016/j.micpro.2018.12.002
[29] Revanna D, Anjum O, Cucchi M, Airoldi R, Nurmi J. A scalable FFT processor architecture for OFDM based communication systems. In: International Conference on Embedded Computer Systems: Architectures, Modeling and Simulation; Agios Konstantinos, Greece; 2013. pp. 19-27.
[30] Abdoli H, Nikmehr H. A flexible CO-OFDM using reconfigurable multi-precision FFT. IEEE Communications Letters 2017; 21: 1997-2000. doi: 10.1109/LCOMM.2017.2706675
[31] Glittas AX, Sellathurai M, Lakshminarayanan G. A normal I/O order radix-2 FFT architecture to process twin data streams for MIMO. IEEE Transactions on Very Large Scale Integration Systems 2016; 24 (6): 2402-2406. doi: 10.1109/TVLSI.2015.2504391
[32] Changela A, Zaveri M, Verma D. FPGA implementation of high-performance, resource-efficient radix-16 CORDIC rotator based FFT algorithm. Integration 2020; 73: 89-100. doi: 10.1016/j.vlsi.2020.03.008
[33] Wang J, Xie Y, Li B, Yang C, Hu S. The reconfigurable pipelined variable-point FFT processor design. In: IEEE International Conference on Signal, Information and Data Processing; Chongqing, China; 2019. pp. 1-4.