• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 98 subscribers
  • Views 383 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.

dm6467 using TMS320C62x/64x FastRTS Library

zhiqiang li
Intellectual 270 points

Hello everyone,

I want to use TMS320C62x/64x FastRTS Library on dm6467, but I don't know how to install it in the codec engine and the detail steps of  using it on davinci platforms?

could anybody help me?

thank you!

zhiqiang

  • 0
    TI__Guru** 116960 points

    Zhiqiang,

    The TMS320C62x/64x FastRTS library can be accessed through the codec engine from the ARM side on DM6467 using the C6Accel package which is available here.

    http://focus.ti.com/docs/toolsw/folders/print/c6accel-dsplibs.html?DCMP=DSP-C6RunAccel-092110&HQS=Other+PR+c6accel-prtf.

    Instructions for using C6Accel on DM6467 are archived here

    http://processors.wiki.ti.com/index.php/Using_C6Accel_on_DM6467_with_DVSDK_3.x.

    Just as a pointer I would like to inform you that using C6Accel to access fastRTS functions on the DSP is beneficial only when the size of the data is large due to the overhead involved in ARM calling functionality on the DSP. Hence you must consider these overheads while making performance estimates.

    Regards,

    Rahul

     

  • 0 in reply to Rahul Prabhu
    Intellectual 270 points

    Hi Rahul,

    You say that 

     "using C6Accel to access fastRTS functions on the DSP is beneficial only when the size of the data is large due to the overhead involved in ARM calling functionality on the DSP", can you explain it in details, I don't quite understand it.

    Thank you!

    Zhiqiang

  • 0 in reply to zhiqiang li
    TI__Guru** 116960 points

    Hi Zhiqiang,

    C6Accel calls functionality on the DSP by making calls through the codec engine. ARM calling a functionality on the DSP  via the codec engine involves the over head of Cache management and address translations which to an extent offset the performance we achieve on the DSP for small values of data. But when large data sizes need to be processed (typically over 1K  bytes) needs to be processed the DSP performance is better despite these overheads and so it becomes beneficial to offload code to the DSP.

    Please take a look at this section http://processors.wiki.ti.com/index.php/C6Accel_ARM_user_Documentation#C6Accel_Overhead for more details.

    Regards,

    Rahul