TMS320C6748: FFT function

Dylan_hua

Part Number: TMS320C6748
Other Parts Discussed in Thread: FFTLIB

Hi,

I use TI DSPF_sp_fftSPxSP function,why maximum value of the parameter N is 131072?

void DSPF_sp_fftSPxSP (int N, float *ptr_x, float *ptr_w, float *ptr_y, unsigned char *brev, int n_min, int offset, int n_max)

--------------------------------------------------------------------------------

Function Documentation
void DSPF_sp_fftSPxSP ( int N,
float * ptr_x,
float * ptr_w,
float * ptr_y,
unsigned char * brev,
int n_min,
int offset,
int n_max
)

over 7 years ago

0 Yordan Kovachev over 7 years ago

TI__Guru**** 161600 points

Which Processor SDK RTOS version is this?

Best Regards,
Yordan

0 Dylan_hua over 7 years ago in reply to Yordan Kovachev

Prodigy 50 points

I don’t know the version, but it is a lib file from TI DSPlib.

0 Rahul Prabhu over 7 years ago

TI__Guru** 116770 points

Dylan,

If you look at the documentation of the API, it is optimized with the assumption that N is power of 2 and 8 <= N <= 131072. The assembly optimization of this has been done years back so it will not be feasible for us to provide exact motivation/limitation to cap this at 131072.

However, we support larger FFT computation using VLFFT provided in FFTLIB but I suspect that this offering is only available for the C66x DSP and not for the C674x DSP.

Regards,

Rahul

Processors

Processors forum

TMS320C6748: FFT function