A library for computing the Discrete Fourier Transform; targeting Graphics Processing Units; supporting OpenCL, Level Zero, and SYCL; with double-batching.
Introduction • Features • Installation • Documentation • License
The Double-Batched FFT Library is a library for computing the Fast Fourier Transform (FFT) on Graphics Processing Units (GPUs). GPU support is enabled via SYCL, OpenCL, or Level Zero.
A distinctive feature is the support of double-batching. That is, given the M x N_1 x ... x N_d x K input tensor, where the Fourier transform shall be taken over the N-modes, the library support batching over the M-mode and K-mode in a single kernel. Single-batching, offered by many FFT libraries, is included as the special case M=1.
- Forward and backward FFTs with complex input data and complex output data (c2c)
- FFTs with real input data and complex output data (r2c)
- FFTs with complex input data and real output data (c2r)
- 1d, 2d, and 3d FFTs
- Single and double precision
- Single-batching and double-batching
- User callbacks written in OpenCL-C for loads and stores (only in 1d)
- Optimized for small FFTs with N <= 512
Install the oneAPI Base Toolkit. Build and install the project with
cmake -Bbuild -S. -DCMAKE_CXX_COMPILER=icpx -DCMAKE_BUILD_TYPE=Release
cmake --build build
cmake --install build --prefix /path/to/installationThe documentation is available at https://intel.github.io/double-batched-fft-library/
Install Doxygen and Python dependencies:
apt install doxygen
pip install -r docs/requirements.txtBuild and install with CMake (add -DBUILD_DOCUMENTATION=ON). To read the docs:
cd /path/to/installation/share/doc/bbfft
python -m http.server 8000Open http://127.0.0.1:8000/ in a browser.