This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

SECDEVTOOL-OMAPL138C6748: Test of comlex FFT and real FFT operation time

Part Number: SECDEVTOOL-OMAPL138C6748
Other Parts Discussed in Thread: OMAP-L138

Hi,

When I test the time difference between 4096 point complex FFT and real FFT, given at the following page, we found two weird phenomenons. 

/cfs-file/__key/communityserver-discussions-components-files/791/2158.Real_5F00_fft_5F00_demo_5F00_c674x.zip

First,  both  the cycle counts of the FFT clock is zero, the following message is given, "Clocks for 4096 Real FFT using N/2 Complex FFT= 0". It  seems the clock() does not work.

Sencond, because the clock does not work, I flip GPIO to get the complex FFT operation time, corresponding code is given below,

    while(1)
    {
        GPIOPinWrite(SOC_GPIO_0_REGS, 9, GPIO_PIN_HIGH);
        DSPF_sp_fftSPxSP (N, pCFFT_In, twiddle, pCFFT_Out, brev, rad, 0, N);
        GPIOPinWrite(SOC_GPIO_0_REGS, 9, GPIO_PIN_LOW);
        DSPF_sp_fftSPxSP (N, pCFFT_In, twiddle, pCFFT_Out, brev, rad, 0, N);
    }

However,I found the duty cycle is not equal, and the operation time of complex FFT is only 450 us.  GPIO is not used anywhere else. The same happens to real FFT for 920us. Plus, I am wondering why the complex FFT is faster than real FFT.

Your quick reply will be highly appreciated.

Regards,

wong