Integer FFT/IFFT cores

This project contains fully pipelined integer unscaled and scaled (truncated LSB) FFT/IFFT cores for FPGA, Scheme: Radix-2, Decimation in frequency and decimation in time;

Integer data type and twiddles with configurable data width.

Code language - VHDL, Verilog

Vendor: Xilinx, 6/7-series, Ultrascale, Ultrascale+;

Smallest FPGA resourses and highest processing frequency that you ever seen!

License: GNU GPL 3.0.

Main information

Title	Universal integer FFT cores (Xilinx FPGAs)
Author	Alexander Kapitanov
Project	Habrahabr
Project lang	VHDL, Verilog
Vendor	Xilinx: 6/7-series, Ultrascale, US+
Release Date	13 May 2018
Last Update	11 Jan 2019

List of complements:

• FFTs:
  • int_fftNk – main core - Full-precision or Scaled FFT, Radix-2, DIF, input flow - normal, output flow - bit-reversed.
  • int_ifftNk – main core - Full-precision or Scaled IFFT, Radix-2, DIT, input flow - bit-reversed, output flow - normal.
• Butterflies:
  • int_dif2_fly – Full-precision or Scaled butterfly Radix-2, decimation in frequency,
  • int_dit2_fly – Full-precision or Scaled butterfly Radix-2, decimation in time,
• Complex multipliers:
  • int_cmult_dsp48 – main integer complex multiplier contains several cmults:
    • int_cmult18x25_dsp48 – simple 25 x 18 two’s-complement half-complex-multiplier,
    • int_cmult_dbl18_dsp48 – double 42(44) x 18 two’s-complement half-complex-multiplier,
    • int_cmult_dbl35_dsp48 – double 25(27) x 35 two’s-complement half-complex-multiplier,
    • int_cmult_trpl18_dsp48 – triple 59(61) x 18 two’s-complement half-complex-multiplier,
    • int_cmult_trpl52_dsp48 – triple 25(27) x 52 two’s-complement half-complex-multiplier,

"half" means that you should set output flow: Re or Im part.

• Multipliers:

  • mlt42x18_dsp48e1 – 42 x 18 two’s-complement multiplier (DSP48E1), del.: 4 taps, res.: 2 DSPs.
  • mlt59x18_dsp48e1 – 59 x 18 two’s-complement multiplier (DSP48E1), del.: 5 taps, res.: 3 DSPs.
  • mlt35x25_dsp48e1 – 35 x 25 two’s-complement multiplier (DSP48E1), del.: 4 taps, res.: 2 DSPs.
  • mlt52x25_dsp48e1 – 52 x 25 two’s-complement multiplier (DSP48E1), del.: 5 taps, res.: 3 DSPs.
  • mlt44x18_dsp48e2 – 44 x 18 two’s-complement multiplier (DSP48E2), del.: 4 taps, res.: 2 DSPs.
  • mlt61x18_dsp48e2 – 61 x 18 two’s-complement multiplier (DSP48E2), del.: 5 taps, res.: 3 DSPs.
  • mlt35x27_dsp48e2 – 35 x 27 two’s-complement multiplier (DSP48E2), del.: 4 taps, res.: 2 DSPs.
  • mlt52x27_dsp48e2 – 52 x 27 two’s-complement multiplier (DSP48E2), del.: 5 taps, res.: 3 DSPs.

• Adder:

  • int_addsub_dsp48 – based on DSP48, up to 96-bit two’s-complement addition/substraction.

• Delay line:

  • int_delay_line – main delay line, cross-commutation data between butterflies.
  • int_align_fft – data and twiddle factor alignment for butterflies in FFT core,
  • int_align_fft – data and twiddle factor alignment for butterflies in IFFT core,

• Twiddles:

  • rom_twiddle_int – 1/4-periodic signal, twiddle factor generator based on memory and sometimes uses DSP48 units for large FFTs
  • row_twiddle_tay – twiddle factor generator which used Taylor scheme for calculation twiddles.

• Buffers:

  • inbuf_half_path – simple input buffer, perform flow into two flows: [0 .. NFFT/2), (NFFT/2 .. NFFT-1], single clock,

  • outbuf_half_path – simple input buffer, merge two flows into one signal, single clock,

  • iobuf_flow_int2 – Mode-1: BITREV = FALSE: convert Interleave-2 flow into two parts of input flows, Mode-2: BITREV = TRUE: convert two-half flows into Interleave-2 signal.

  • int_bitrev_ord – simple converter data from bit-reverse to natural order.

Author:

• Kapitanov Alexander

Release:

• 2018/13/05

License:

• GNU GPL 3.0