• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 83 subscribers
  • Views 850 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

CCS/TMS320F280049C: FFT algorithm execution using the 280049 CLA and 280049 floating point capabilities.

UserHomeInit
Expert 2145 points
Part Number: TMS320F280049C
Other Parts Discussed in Thread: TMS320F280049, C2000WARE

Tool/software: Code Composer Studio

Dear TI Hello,

I am working on the TMS320F280049 DSP and i have questions about the c28x floating core and the Control Law Accelerator (CLA) capabilities.

1. Does the TMS320F280049 contains any new specific method (feature/peripheral) or compiler options to help me execute faster floating operations and FFT algorithm respectively.

2.It is possible to execute the FFT algorithm using the 280049  CLA ?

3.Does the TI RFFT C macro  could be executed on the 280049 CLA? 

4.Does TI has any examples for FFT in the resource library CLA examples?  

Thank you in advance,

Regards,

S.Tarik,

  • 0
    TI__Mastermind 25030 points

    Tarik,

    1. You can refer to the F28004x datasheet for details. The TMU should help speed up floating point operations, but this is not something newly added. Compiler updates keep occurring, somewhat independent of device development.

    2. Yes, it would not be as efficient as implementing it on the C28x CPU, but if you absolutely need to do it for some reason, you can of course do it.

    3. Not sure which code you are referring to - if it is purely C, then of course it can be run on the CLA. A related comment is that the CLA C compiler now allows nesting functions (but not nested CLA tasks). If it is a mix of C and C28x asm, then that would not work as the asm would need to be rewritten for the CLA.

    4. I don't believe we have anything in C2000Ware. But we have something here that might be useful. You can download the project files under "Additional Information".

    training.ti.com/cla-hands-workshop-part-3-workshop

    Finally, if you have time on your hands, you can refer to this related discussion on e2e

    https://e2e.ti.com/support/microcontrollers/c2000/f/171/t/758021?TMS320F28379D-Purpose-of-CLA-FFT-Table

    Thanks,

    Sira

  • 0 in reply to Sira Rao80
    Expert 2145 points

    Hello Sira,

    I really appreciate your respond and the training link,

    The FFT algorithm is an .asm function, RFFT_f32.asm. Could you tell me please what do you mean by purely C.

    The files I would like to  execute on the CLA to compute the FFT algorithm are:

    RFFT_f32_mag.asm

    RFFT_f32_phase.asm

    RFFT_f32_sincostable.c

    RFFT_f32.asm

    These functions are written in assembly and used in C code. Do you think is a mix of C and C28x asm?

    Thank you in advance,

    S.Tarik

  • 0 in reply to UserHomeInit
    TI__Mastermind 25030 points

    Tarik,

    Yes, the core algorithm is implemented in C28x assembly, which is why it can't be ported as is to the CLA because the CLA instruction set is different.

    Thanks,

    Sira

  • 0 in reply to Sira Rao80
    Expert 2145 points

    Thank you Sira,