rust-fftw3

Rust bindings for the FFTW C-library for computing discrete Fourier transforms, as well as discrete cosine and sine transforms.

This repository includes three crates:

• fftw: A safe wrapper in Rust
• fftw-sys: An unsafe wrapper in Rust
• fftw-src: A crate for downloading and compiling the FFTW library

Feature flags

• source: Download and compile FFTW (default)
  • (Linux, macOS) Needs a C-compiler and the make build tool to compile the FFTW library
  • (Windows) Downloads a precompiled binary from the FFTW website
• system: Use the system's libfftw3 (experimental)
  • You must install FFTW before building this crate
  • For Linux systems, please install FFTW using your package manager, e.g. in Debian or Ubuntu run apt install libfftw3-dev
  • For macOS, please run brew install fftw by using homebrew
  • This feature is unsupported on Windows
• intel-mkl Use Intel MKL backend through intel-mkl-src
  • Only Linux and Windows are supported

LICENSE

See LICENSE.md

lib.rs:

Rust binding of FFTW

Examples

Complex-to-Complex

use fftw::array::AlignedVec;
use fftw::plan::*;
use fftw::types::*;
use std::f64::consts::PI;

let n = 128;
let mut plan: C2CPlan64 = C2CPlan::aligned(&[n], Sign::Forward, Flag::MEASURE).unwrap();
let mut a = AlignedVec::new(n);
let mut b = AlignedVec::new(n);
let k0 = 2.0 * PI / n as f64;
for i in 0..n {
    a[i] = c64::new((k0 * i as f64).cos(), 0.0);
}
plan.c2c(&mut a, &mut b).unwrap();

Complex-to-Real

use fftw::array::AlignedVec;
use fftw::plan::*;
use fftw::types::*;
use std::f64::consts::PI;

let n = 128;
let mut c2r: C2RPlan64 = C2RPlan::aligned(&[n], Flag::MEASURE).unwrap();
let mut a = AlignedVec::new(n / 2 + 1);
let mut b = AlignedVec::new(n);
for i in 0..(n / 2 + 1) {
    a[i] = c64::new(1.0, 0.0);
}
c2r.c2r(&mut a, &mut b).unwrap();