Project Release Information
This release adds ARM NEON support, implements faster estimate-mode planning, and improves Fortran interface portability.
A finalized release of AVX, MPI, and other new features in 3.3-beta1. Compared to the beta, the main change is support for simultaneous installation of OpenMP and POSIX-threads versions.
Support was added for AVX instructions, MPI distributed-memory parallelism, Fortran 2003 interface, quadruple precision (gcc 4.6 or later), and several minor enhancements.
Some performance regressions were fixed. A workaround was added for a mingw64 bug.
Performance improvements were in multidimensional real-data transforms. Documentation and compilation fixes were made.
FFTW is a fast C FFT library. It includes complex, real, symmetric, multidimensional, and parallel transforms, and can handle arbitrary array sizes efficiently.It is typically faster than other freely available FFT implementations, and is even competitive with vendor-tuned libraries (benchmarks are available at the homepage). To achieve this performance, it uses novel code generation and runtime self optimization techniques (along with many other tricks).