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/LAUNCHXL-F28379D: Time consuming of icfft/cfft and linear fitting library functions

Part Number: LAUNCHXL-F28379D
Other Parts Discussed in Thread: C2000WARE

Tool/software: Code Composer Studio

Dear colleagues,

1. I want to use a set of data points to fit a straight line. This work needs to be completed on the F28379D.

Is there any library functions for linear fitting?

2. I found the example project 2837x_cfft in C2000Ware.

How much time is consumed by different points of ICFFT/CFFT operation? Such as 128, 256, 512, 1024, 2048 points.

Best regards,

Hao

  • Dear Hao,

    Thanks for your questions.

    Unfortunately there are currently no C2000 library routines for linear curve fitting.

    The execution benchmarks for the cfft_32t function can be found in Table 5 on p.28 of the user's guide for the FPU32 DSP software lib. If you have installed the latest C2000Ware in the default location, you can find the document at:
    C:\ti\c2000\C2000Ware_1_00_04_00\docs

    The inverse operation benchmarks are on p.45 of the same document.

    Please post back if anything's unclear.  Thanks again.

    Regards,

    Richard

  • Hi Richard,

    Thank you for your help!

    I use two CPUs to program.

    If the project of CPU1 want to use an matrix which is defined in the project of CPU2, what should I do?

    Best regards,

    Hao

  • Hi Hao,

    Do you mean a #define shared between two projects, or passing data between the CPUs? If the former, couldn't you include the same header in both projects?

    To share data between CPUs you have two options:
    - load the data to global shared RAM blocks, or
    - use the IPC module.

    In the first case, CPU2 would write the matrix into the appropriate GS memory block which CPU1 could read. You configure write access using the GSxMSEL register.

    In the latter case, CPU2 would write the matrix into IPC message RAM and use the IPC flags to signal to CPU1 that the data is ready. I suggest this is the better method since you can synchronize data accesses.

    Both methods are covered nicely in module 11 of the multi-day training course for this device, which you can access at:
    processors.wiki.ti.com/.../C2000_Multi-Day_Workshop

    Regards,

    Richard
  • Hi Richard,

    I am sorry I didn't express clearly.

    I'll try the latter methods: using the IPC module.

    Thank you!

    Best regards,

    Hao