Since JP6 nvidia-smi is now working on Orin AGX but it does not show any sensible data like GPU usage. No process is shown and Usage always stays at 0 %.
As this is the de facto standard to measure CPU usage on Nvidia this is a major problem for Algorithm development
Hi,
Here are some suggestions for the common issues:
Please run the below command before benchmarking deep learning use case:
$ sudo nvpmodel -m 0
$ sudo jetson_clocks
Installation guide of deep learning frameworks on Jetson:
Startup deep learning tutorial:
If these suggestions don’t help and you want to report an issue to us, please attach the model, command/step, and the customized app (if any) with us to reproduce locally.
Thanks!
@carolyuu THX but those tip do not apply to my issue.
Please try with Tegrastats: Tegrastats Utility — NVIDIA Jetson Linux Developer Guide 1 documentation
@kayccc
Thx we are using that already. But as we have different teams and different algorithms on the same hardware we need also per process numbers which as far as I know tegrastats can not deliver.
As nvidia-smi is part of JP6 I assume it should work and this is a bug.
Hi,
Thanks for reporting this.
Confirmed that we can reproduce similar behavior in our environment.
We need to check with our internal team for more info.
Will provide more info to you.
Thanks.
@AstaLLL were you able to get any info from the team?
Hi,
Not yet, will keep you updated.
Thanks.
Hi,
This is a feature request.
Currently, nvidia-smi functionality is quite limited on Jetson.
The GPU utilization and the processes using GPU are not yet supported.
We have filed a feature request for this.
It will be evaluated and prioritized based on our internal resources.
Thanks.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.