Usage of DLA Cores running Holoscan app

Hi, there is any possibility of using Jetson Orin Devkit DLA cores during Holoscan’s TensorRT inference?

I checked Jtop and DLA cores are not used by default. If i use them i can have an improve of performance. Is there any documentation on how to do this?

Thanks

Hi Alfonso,

Sorry for the late response.
Currently the Holoscan SDK inference operator does not support usage of the DLA. But we are working on exposing that feature on a future release.

Regards,
Andreas