Hello there,
This might be a silly question, so pardon me if it is.
Let us say that I want to perform M 1D DFTs (let us say, complex-complex) on an array of dimension N, but I only need the first P entries instead of M (e.g. by construction, I know that only the first P entries will be nonzero).
Right now what I do is to compute it full and then discard the last (M-P) entries. It then requires 2 arrays, one temporary, MxN, and the truncated one which I'll effectively use in my code, PxN. Also, if I want to do the backward transform I need to copy the PxN array into an MxN array, padded with zeroes, and then call the transform.
Is there any way to avoid the MxN array altogether and tell MKL that I am just interested in the first P entries for each 1D DFT?