How to modify the official website's fft1d kernel to get a 32k point Fourier transfor
I met a problem that when I modified the official kernel with a larger points than 4k, I can not get the correct result. How can I modify the code for a large number fft?