falling back to GPU

Autonomous Vehicles DRIVE AGX Xavier DRIVE AGX Xavier General
1

Dear.
When I ran xavier with “./trtexec --avgRuns=100 --deploy=./ResNet-50-deploy.prototxt --fp16 --batch=8 --iterations=10 --output=prob --useDLACore=1 --useSpinWait --allowGPUFallback”
I got below message.
Default DLA is enabled but layer prob is not running on DLA, falling back to GPU.

Is this mean that only ‘prob’ layer in the resnet50 run on GPU and the others on NVDLA? all layers on GPU?

and I want to know how to monitor xavier. is there any tool to it?

2

Dear rooney.sm.yang,
Please check https://docs.nvidia.com/deeplearning/sdk/tensorrt-developer-guide/index.html#trtexec . All the DLA supported layers runs on DLA and rest all gets execute on GPU.

3

Dear SivaRamaKrishna.
do you mean this?
All the DLA supported layers runs on DLA and rest all gets execute on GPU.

4

Yes. I have corrected the typo. Thank you.

5

Dear SivaRamaKrishna.

Thank for your reply.