Hello, I am currently attempting to build a TensorRT engine to utilise the DLA cores on a Jetson AGX Orin Development Kit. When trying to pass layers to the DLA I am met with there not being enough local DRAM to store the layers and the output says “Hint: Try increasing the DLA memory pool limits”. My question is how exactly do I do this? I can’t find any docs anywhere describing increasing these memory limits.
As a follow up, if we increase the DLA memory pool, does this subtract memory from the rest of the board?