| Re: [eigen] Question on eigen/FFT |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
Am 09.03.2016 um 15:55 schrieb Mark Borgerding:
Hi.
I am the original author of the unsupported FFT module. Unfortunately,
I don't have the time now to support it and no one else has stepped up
and volunteered to take the reigns.
One thing regarding completenes:
I am missing something as fftshift (in matlab)
I have my own implementation, but would like to share it,
because I have little experience with best coding practive of
functions basically shifting around memory .
re alignment.
The Eigen allocation should be sufficient.
I assumed that, but although I do like Eigen, I do not want to depend on it.
If you wish to use std::vector, then you'll probably want to
specialize it with an aligned_allocator -- a quick google search
should turn something up there.
Indeed, but all I found code, but they all make use of features I do
know little of or can not test, some even depend on C++11.
http://stackoverflow.com/questions/12942548/making-stdvector-allocate-aligned-memory
http://jmabille.github.io/blog/2014/12/06/aligned-memory-allocator/
I would rather want to depend on something tested and proven.
Boost provides something similar:
http://www.boost.org/doc/libs/1_60_0/doc/html/align/examples.html#align.examples.aligned_vector
But I do not know if that gives me what I need for fftw
The fall back would be to use C-Pointers and the fftw malloc functions.
The return-by-value proxy is an Eigen-ism that allows expressions to
be composed and simplified before evaluation. The FFT module does not
exploit this power well, if at all.
I removed the code since I do not need it.
Matthias