FFTW++: 3.01 Released

2024-01-09T08:17:38.268000Z

FFTW++ provides a simple C++ interface for 1D, 2D, and 3D complex-to-complex, real-to-complex, and complex-to-real Fast Fourier Transforms that takes care of the technical aspects of memory allocation, alignment, planning, wisdom, and communication on both serial and parallel (OpenMP/MPI) architectures. Wrappers for multiple 1D transforms are also provided. As with the FFTW3 library, in-place and out-of-place multithreaded transforms of arbitrary size are supported. Hybrid dealiasing of standard and centered Hermitian convolutions is also implemented; implicit zero-padding substantially reduces memory usage and computation time. Parallelization is accomplished using an efficient adaptive OpenMP/MPI transpose routine. Wrappers for C, Python, and Fortran are included.

FFTW++: 2.02 Released

2016-02-20T08:05:42.979000Z

FFTW++ provides a simple C++ interface for 1D, 2D, and 3D complex-to-complex, real-to-complex, and complex-to-real Fast Fourier Transforms that takes care of the technical aspects of memory allocation, alignment, planning, wisdom, and communication on both serial and parallel (OpenMP/MPI) architectures. Wrappers for multiple 1D transforms are also provided. As with the FFTW3 library, in-place and out-of-place multithreaded transforms of arbitrary size are supported. Implicit dealiasing of standard and centered Hermitian convolutions is also implemented; in 2D and 3D this implicit zero-padding substantially reduces memory usage and computation time. Parallelization is accomplished using an efficient adaptive hybrid OpenMP/MPI transpose routine. Wrappers for C, Python, and Fortran are included.

An out-of-place optimization was enabled and potential premature string deallocation bugs were fixed.

FTW++: 2.01 Released

2016-02-06T08:01:26.669000Z

FFTW++ provides a simple C++ interface for 1D, 2D, and 3D complex-to-complex, real-to-complex, and complex-to-real Fast Fourier Transforms that takes care of the technical aspects of memory allocation, alignment, planning, wisdom, and communication on both serial and parallel (OpenMP/MPI) architectures. Wrappers for multiple 1D transforms are also provided. As with the FFTW3 library, in-place and out-of-place multithreaded transforms of arbitrary size are supported. Implicit dealiasing of standard and centered Hermitian convolutions is also implemented; in 2D and 3D this implicit zero-padding substantially reduces memory usage and computation time. Parallelization is accomplished using an efficient adaptive hybrid OpenMP/MPI transpose routine. Wrappers for C, Python, and Fortran are included.

Deadlocks in the latency calculation were fixed, along with two memory leaks. Timing scripts were improved.

FTW++: 2.00 Released

2015-11-09T00:47:40.381000Z

FFTW++ provides a simple C++ interface for 1D, 2D, and 3D complex-to-complex, real-to-complex, and complex-to-real Fast Fourier Transforms that takes care of the technical aspects of memory allocation, alignment, planning, wisdom, and communication on both serial and parallel (OpenMP/MPI) architectures. Wrappers for multiple 1D transforms are also provided. As with the FFTW3 library, in-place and out-of-place multithreaded transforms of arbitrary size are supported. Implicit dealiasing of standard and centered Hermitian convolutions is also implemented; in 2D and 3D this implicit zero-padding substantially reduces memory usage and computation time. Parallelization is accomplished using an efficient adaptive hybrid OpenMP/MPI transpose routine. Wrappers for C, Python, and Fortran are included.

FFTW++: 1.13 Released

2012-10-28T18:18:38.120000Z

FFTW++ provides a simple C++ interface for 1D, 2D, and 3D complex-to-complex, real-to-complex, and complex-to-real Fast Fourier Transforms that takes care of the technical aspects of memory allocation, alignment, planning, and wisdom. Wrappers for multiple 1D transforms are also provided. As with the FFTW3 library, in-place and out-of-place multithreaded transforms of arbitrary size are supported. Implicit dealiasing of standard and centered Hermitian convolutions is also implemented; in 2D and 3D this implicit zero-padding substantially reduces memory usage and computation time. Wrappers and examples for C, Python, and Fortran were added. Optimized ternary convolutions were implemented. Multithreaded initialization was improved.

FFTW++: 1.07 Released

2012-10-28T15:32:53.011000Z

FFTW++ provides a simple C++ interface for 1D, 2D, and 3D complex-to-complex, real-to-complex, and complex-to-real Fast Fourier Transforms that takes care of the technical aspects of memory allocation, alignment, planning, and wisdom. Wrappers for multiple 1D transforms are also provided. As with the FFTW3 library, in-place and out-of-place transforms of arbitrary size are supported. Implicitly dealiasing of standard and centered Hermitian convolutions is also implemented; in 2D and 3D this implicit zero-padding substantially reduces memory usage and computation time. Workarounds were implemented for several C99 features not yet implemented in the Microsoft Windows operating system.

FFTW++: 1.06 Released

2012-10-28T15:32:52.963000Z

FFTW++ is a C++ header class for the FFTW Fast Fourier Transform library that automates memory allocation, alignment, planning, and wisdom. In 2D and 3D, implicit dealiasing of convolutions substantially reduces memory usage and computation time.

FFTW++: 1.08 Released

2012-10-28T15:32:52.910000Z

FFTW++ provides a simple C++ interface for 1D, 2D, and 3D complex-to-complex, real-to-complex, and complex-to-real Fast Fourier Transforms that takes care of the technical aspects of memory allocation, alignment, planning, and wisdom. Wrappers for multiple 1D transforms are also provided. As with the FFTW3 library, in-place and out-of-place transforms of arbitrary size are supported. Implicit dealiasing of standard and centered Hermitian convolutions is also implemented; in 2D and 3D this implicit zero-padding substantially reduces memory usage and computation time. The M_PI constant was replaced by a static constant for portability. Timing graphs were improved. Biconvolutions are now called ternary convolutions.

FFTW++: 1.11 Released

2012-10-28T15:32:52.855000Z

FFTW++ provides a simple C++ interface for 1D, 2D, and 3D complex-to-complex, real-to-complex, and complex-to-real Fast Fourier Transforms that takes care of the technical aspects of memory allocation, alignment, planning, and wisdom. Wrappers for multiple 1D transforms are also provided. As with the FFTW3 library, in-place and out-of-place transforms of arbitrary size are supported. Implicit dealiasing of standard and centered Hermitian convolutions is also implemented; in 2D and 3D this implicit zero-padding substantially reduces memory usage and computation time. Support was added for multithreaded transforms and convolutions.

FFTW++: 1.12 Released

2012-10-28T15:32:52.745000Z

FFTW++ provides a simple C++ interface for 1D, 2D, and 3D complex-to-complex, real-to-complex, and complex-to-real Fast Fourier Transforms that takes care of the technical aspects of memory allocation, alignment, planning, and wisdom. Wrappers for multiple 1D transforms are also provided. As with the FFTW3 library, in-place and out-of-place multithreaded transforms of arbitrary size are supported. Implicit dealiasing of standard and centered Hermitian convolutions is also implemented; in 2D and 3D this implicit zero-padding substantially reduces memory usage and computation time. A parallelized loop increment was localized. The SINGLE_THREAD preprocessor symbol was renamed to FFTWPP_SINGLE_THREAD. Examples and tests were organized into separate directories. A doxygen configuration file was added.

Recent posts to news