FFTW is a comprehensive collection of fast C routines for computing the Discrete Fourier Transform (DFT) and various special cases thereof, copyrighted by MIT and distributed under the GNU General Public License. For more information, refer to FFTW website (fftw.org).
FFTW can compute transforms of real and complex-values arrays of arbitrary size and dimension. AOCL-FFTW is an AMD optimized version of FFTW implementation targeted for AMD EPYC CPUs.
Highlights of AOCL-FFTW 3.2
- Upgraded AOCL-FFTW to align with the open source master FFTW 3.3.10 from MIT
- Dynamic dispatcher for AOCL-FFTW
- Windows FFTW features aligned with Linux FFTW
Highlights of AMD FFTW 3.1
- Feature ‘AMD application optimization layer’ that uplifts the performance of HPC and scientific applications
- Feature ‘Fast MPI transpose algorithm’ to speed up the distributed MPI FFT computations
- Feature ‘Top N planner’ that minimizes single-threaded run-to-run variations
- Support for building AOCL-FFTW library on Windows
- GCC compilation support for AMD processors based on the AMD “Zen3” core architecture
The packages containing AOCL-FFTW binaries, examples and documentation are available in the Downloads section below.
Source code for AOCL-FFTW is available on GitHub (https://github.com/amd/amd-fftw).
For prior versions of AOCL-FFTW documentation and downloads, refer to AOCL-FFTW Archive.