topic Re: 2D DFTF using real_2d_ccs_double_ex2.f90 in Intel® oneAPI Math Kernel Library

2D DFTF using real_2d_ccs_double_ex2.f90

nandan_tandongmail_c — Fri, 24 Apr 2009 10:49:06 GMT

Hello,

I want to use the "real_2d_ccs_double_ex2.f90" example, suitably modified to evaluate the FFT
of a 100x100 matrix. I have modified the N_MAX and M_MAX values to 100 and also read the 100X100
matrix into the X_IN_2D matrix. I do not modify any other value and allow the example to run.

The problem i face is that the results for the forward and backward FFT which are seen in "real_2d_ccs_double_ex2.res"
show only 200 elements and writen in a 100x3 (=300) format. I am not able to change the format into a more suitable one.

Also if i understand correctly, the "strides" values will get set once the "n, m " values are appropriately set and they do
not have to be modified directly.

Please let me know whether this is correct.

Regards,

Nandan.

Re: 2D DFTF using real_2d_ccs_double_ex2.f90

Dmitry_B_Intel — Fri, 24 Apr 2009 11:23:49 GMT

Hi Nandan,

The examplesintentionallyprint only a few columns of whole data set. If you need the whole set, you could employthe equivalencesX_OUT<->X_OUT_2Dand X_IN<->X_IN_2D and print the 2D arrays instead of computing indexing into the 1D arrays. However,the 2D CCS format is not easy to interpret, in the sense that real and imaginary parts of the result of the forward transform are dispersed within the real X_OUT_2D array in an irregular way. So I would encourage you to use more regular 2D CCE format instead. In the CCE format, X_OUT_2D(i,j) is exactly the complex number you'd need in the frequency domain, except thati runs only up to M/2+1 (the other half can be restored by conjugate-even symmetry).

Thanks
Dima