• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 98 subscribers
  • Views 128 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.

TDA4VH-Q1: How to Run Post-Processing on C7 Instead of A72 for Semantic Segmentation?

Minho Park
Prodigy 40 points
Part Number: TDA4VH-Q1
Other Parts Discussed in Thread: TDA4VH

Tool/software:

Hi,

I’m currently working on deploying a semantic segmentation model on the TDA4VH board.
At the moment:

The main model graph is executed on the C7 DSP,

But the post-processing stage (e.g., class remapping, smoothing, overlay) is running on the MPU (A72).

For performance reasons, I would like to move the post-processing part to C7 as well.

Could you please advise:

How can I configure or modify the graph to execute the post-processing kernels on the C7 DSP?

Are there specific steps needed for kernel registration, target assignment, or graph setup?

I’m using SDK version vh0902 and building the graph using OpenVX APIs.

Any guidance would be greatly appreciated!

Best regards,
Minho

  • 0
    TI__Expert 7061 points

    Hello;

    Thanks for the question. 

    In general, we don't recommend customers manually modify/config the graphs running on the C7. It could be problematic. 

    But, if you have some specific performance issues of post processing, and if you can provide some details, we will find experts on OpenVX or on other areas to help you.

    Best regards

    Wen Li