I have such a problem - I would need to expand a discrete function in a sine fourier series but I would like to use exponential based library for FFT (I will use CUDA to compute it). What have I to do with the signal I'm sending to the procedure so this is treated as sine fourier? I think I should add zeros but I am not sure where they have to be. Many thanks

  • $\begingroup$ The library you want to use can't directly do a DST? $\endgroup$ – J. M. May 24 '13 at 17:03
  • $\begingroup$ I would like to write the code both for MKL and CUDA. In CUDA, I don't know about Sine Library. And in MKL, the exponential based FFT is available in native MPI parallelization form as well but the sine version not. $\endgroup$ – azbri May 24 '13 at 17:06
  • $\begingroup$ There are at least four DSTs; I presume you want DST-I? $\endgroup$ – J. M. May 24 '13 at 17:09
  • $\begingroup$ yes, DST-I is that $\endgroup$ – azbri May 24 '13 at 17:14
  • 1
    $\begingroup$ Then, have you seen this? $\endgroup$ – J. M. May 24 '13 at 17:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.