# Changes between Version 15 and Version 16 of Examples/Fft2dHighpass

Show
Ignore:
Timestamp:
05/17/10 09:20:03 (6 years ago)
Comment:

--

Unmodified
Removed
Modified
• ## Examples/Fft2dHighpass

v15 v16
4646|| Using FFTW using Estimate mode || 0.09 || [http://code.haskell.org/repa/repa-head/repa-examples/FFT/HighPass/legacy/c/FFTW.c FFTW.c] ||
4747
48 The vector version uses the same recursive radix-2 decimation in time (DIT) algorithm as the Repa version, but is not rank generalised. It applies a recursive 1d FFT to each row and then transposes the matrix, twice. Recursive FFT algorithms tend to be slower than in-place ones because the data is copied into new vectors at each recursion. A 512 point FFT is built from two 256 point FFTs, which are build from 4 128 point FFTs and so on. The result of each FFT is a new vector which needs to be allocated and then filled.
48The vector version uses the same recursive radix-2 decimation in time (DIT) algorithm as the Repa version, but is not rank generalised. It applies a recursive 1d FFT to each row and then transposes the matrix, twice each. Recursive FFT algorithms tend to be slower than in-place ones because the data is copied into new vectors at each recursion. A 512 point FFT is built from two 256 point FFTs, which are build from 4 128 point FFTs and so on. The result of each FFT is a new vector which needs to be allocated, filled, then unboxed again during the next recursion.
4949
50 Jones's version also uses a 1d radix-2 DIT FFT kernel, but it first reorders the values then performs a in-place transform using three nested loops.
50Jones's version also uses a 1d radix-2 DIT FFT kernel, but it first reorders the values then performs a in-place transform using three nested loops. Using an in-place algorithm gives better locality, and avoids the need to allocate and unbox all the intermediate vectors.
5151
5252FFTW contains deep magic, and is comparable with vendor optimised versions.