Re: [Bug-gsl] FFT stride problem

[Brian Gough]

At Wed, 10 Jan 2007 17:40:48 -0500,
Brown, Aaron D wrote:
> I have been getting program crashes after using stride values greater
> than 1 with the FFT functions.  The crashes seem to happen when I try to
> free the data buffer that I passed into the FFT function which I
> dynamically allocated.  My program also crashes sometimes if I free the
> wavetable and workspaces.  Changing the stride value that I pass into
> the FFT function seems to affect which deallocation will fail.  Also, it
> doesn't seem to be limited to a specific FFT function.  The following is
> a test program that will hopefully illustrate the problem that I have
> been having:

Hello,

Thanks for your email.  I don't think this is a problem with the
library.  The input array must be of length 2*n*stride for an n
element complex fft, otherwise you'll overrun the end of the array.

-- 
best regards,

Brian Gough
(GSL Maintainer)

Network Theory Ltd,
Commercial support for GSL --- http://www.network-theory.com/gsl/