kernel m_dftKernels [TP_CASC_LEN *TP_SSR]
The array of kernels that will be created and mapped onto AIE tiles. There will be TP_SSR number of parallel cascade chains of length TP_CASC_LEN .
port <input> in [TP_SSR *TP_CASC_LEN]
The input data to each of the kernels in the function. This input is a window of samples of TT_DATA type. The number of samples in the window to each kernel is derived by (TP_POINT_SIZE / TP_CASC_LEN) * *TP_NUM_FRAMES . A data frame of size TP_POINT_SIZE may require zero-padding so that after it has been divided by TP_CASC_LEN, *so that it is aligned to the vector read size (for AIE this would 4, 8, and 8 samples for cint16, cint32, and cfloat respectively, and for AIE-Ml this would be *8 samples for all data types). Each kernel in a cascaded design will receive an equal share of the input data. Further zero-padding of the input data to each kernel in cascade may be required to ensure this is possible. More information about the required padding of the input data can be found in the documentation for the DFT.
port <output> out [TP_SSR]
The output data of the function. For cascaded designs, this is located at the end of the cascaded kernel chain. This input will be a complex TT_DATA type, and there will be TP_SSR output ports. The number of samples in each SSR output window is derived by (TP_POINT_SIZE / TP_SSR) * TP_NUM_FRAMES .