FFT benchmark

A benchmark for comparison of FFT algorithms performance.

Currently supports Intel IPP, KFR, FFTW and KissFFT.

Requires:

• Clang 6.0+ (GCC and MSVC are not suitable)
• Latest MSYS2 on Windows
• CMake 3.0 or newer
• AVX2-capable cpu
• Python 3.5 or newer
  • matplotlib module
  • numpy module

Usage

Place headers to include folder:

include/
    fft/
        fftw3.h
    ipp/
        ipp.h
        <other IPP headers>
    kfr/
        dft.hpp        
        <other KFR headers from include/kfr>

x64 libs to lib folder:

lib/
    ipps.lib
    <other IPP libs>

x64 dlls to bin folder, including MinGW system dlls:

bin/
    ipps.dll
    <other IPP dlls>
    ...
    libfftw3-3.dll
    libfftw3f-3.dll
    ...
    libstdc++-6.dll
    libwinpthread-1.dll
    libgcc_s_seh-1.dll

Run msys2-build.cmd to build.

Run benchmark.py to measure performance and save data/images.

License

Dual licensed under GPL 2+ and MIT