• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 99 subscribers
  • Views 1608 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Tms320c6678: TMS320C6678

S Siddiqui
Intellectual 385 points
Part Number: TMS320C6678
Other Parts Discussed in Thread: FFTLIB

My goal is to compute 1mega point FFT using c6678.I have used single precision  fft example, available with Dsplib versrion 3.4.Using this example I have computed FFTs of 1K,64K ,128K and 256K points.On comparig th results with matlab I have observed that the FFT lib function work fine for points less than and equal to 64k but for points greater than and equal to 128k results computed from lib function do not match with Matlab.Kindly guide me how I can take 1MEGA point FFT using single core of c6678

Regards

  • 0
    TI__Guru* 84110 points
    Can you offer your test-case code example with input data and expected and resulting outputs for us to compare, please? There are several functions in the library and many ways to write a test, so your test case is the most practical step forward.

    Since you are using the C6678, have you tried using the FFTLIB functions or only the DSPLIB functions?

    Regards,
    RandyP
  • 0 in reply to RandyP
    Intellectual 385 points

    I have tried only following DSPLIB, single precision FFT function "DSPF_sp_fft_SPxSP"

    can you guide me which function of  FFTLIB will be suitable for my case also share sample code pls.

    regards

  • 0 in reply to S Siddiqui
    TI__Guru* 84110 points
    Please download and install the FFTLIB which is for the C66x DSP core. Look through the supplied documentation and let us know if you have questions on anything specific. I cannot judge for your application. And I do not know where the sample routines are, but if you do not find them from the installation, please query back here and someone may be able to help you better than I can.

    Regards,
    RandyP
  • 0 in reply to RandyP
    Intellectual 385 points
    Kindly guide me how to use fft_sp_1d_c2c kernnel of FFTLIB v2.0
  • 0 in reply to S Siddiqui
    Intellectual 385 points
    I would like to know for computing 1Mega point FFT using "fft_sp_1d_c2c_direct" kernnel of FFTLIB V2.0, what values should be assigned to following parameters of the said kernnel

    M Power of 2 FFTSize
    N_P2 Power of 2 FFTSize in mixed radix
    S_r3 No. of stages of radix3
    S_r5 No. of stages of radix5
    Bn Bluestein Pointer
    workbuf Bluestein Scratchpoints
    void *edmastate

    Regards