Hi Kaleb,
Thanks for your right answer.
I want to ask another question about floating point number.
I use CMSIS DSP library to do FFT (fast Fourier transform), this transform will use real number and complex number.
I use "Redlib (nohost-nf)" library now, and it seems ok to compile FFT API like "arm_rfft_fast_f32()".
Do I need to change the "Redlib (nohost-nf)" to another library for using floating point number?