Why my model always runs on DLA 0

[QQQQ]

I generated two trt model through trtexec from the same onnx. with cmd

NVIDIA Jetson AGX Xavier [16GB] - Jetpack 4.6.2 [L4T 32.7.2]
trt 8.2

./trtexec --onnx=model.onnx --saveEngine=model_dla_1_fixshape_ip1.trt --workspace=4096 --verbose --fp16 --minShapes=input_1:0:1x160x160x1 --optShapes=input_1:0:1x160x160x1 --maxShapes=input_1:0:1x160x160x1 --useDLACore=1  --allowGPUFallback

./trtexec --onnx=model.onnx --saveEngine=model_dla_1_fixshape_ip1.trt --workspace=4096 --verbose --fp16 --minShapes=input_1:0:1x160x160x1 --optShapes=input_1:0:1x160x160x1 --maxShapes=input_1:0:1x160x160x1 --useDLACore=0  --allowGPUFallback

and i use the heck status command

$ watch -n 1 "cat /sys/devices/platform/host1x/15880000.nvdla0/power/runtime_status"
$ watch -n 1 "cat /sys/devices/platform/host1x/158c0000.nvdla1/power/runtime_status"

always dla=0  return active and dla=1 return suspend

with the running cmd
./trtexec --loadEngine=model_dla_1_fixshape_ip1.trt --streams=1 --shapes=input_1:0:1x160x160x1 --iterations=1000 --fp16

./trtexec --loadEngine=model_dla_0_fixshape_ip1.trt --streams=1 --shapes=input_1:0:1x160x160x1 --iterations=1000 --fp16

even though i set trtRuntime->setDLACore(1); in cpp code

[AastaLLL]

Hi,

Is JetPack 5 an option for you?
If yes, could you check if the same issue also occurs on the JetPack 5?

Thanks.

[QQQQ]

will retest it after we update to jp 5