Lower power consumption with more input sources in DeepStream

We recently measured the power requirements of a Jetson Nano running a DeepStream-based audio data classification application. Since the number of audio sources is dynamically configured, we measured this with different numbers of sources. The model itself is designed to use up to 10 input sources, some of which are not used if the local audio sources are not available.

With a single audio source, the Jetson requires an average of 3.18 W. With five and ten sources connected, the Jetson requires 3.16 W and 3.13 W, respectively. Particularly noteworthy is the station’s lower power consumption when a larger number of Mics are connected. Also, the GPU utilization is lower with many mics connected compared to smaller number (98% with 1 client, 88% with 5 clients, 80% with 10 clients).

Does anyone have a good explanation for this observed behavior? It is not clear to me how DeepStream handles the missing audio inputs, or what happens in these cases.

Thank you very much,
Jonas

Hi,
Audio decoding is on CPU cores. Please run sudo tegrastats to check and see if there deviation between one source and five or ten sources. The loading is much lower when comparing to video sources, so it probably is not easy to see the difference.

Okay, my measurements were done using tegratstats and a custom wrapper which outputs data as csv. I made both, data and the Jupiter notebook used for looking into the data available here at GitHub: GitHub - umr-ds/jetson-power

The Jupyter Notebook can be viewed using nbviewer: https://nbviewer.org/github/umr-ds/jetson-power/blob/1/jetson.ipynb

As you already suggested CPU usage raises with more audio sources, however GPU utilization and power demand goes down…

Hi,
We would need to reproduce the issue and check. Please share the steps to set up and reproduce the issue. And please share the information:
• Hardware Platform (Jetson / GPU)
• DeepStream Version
• JetPack Version (valid for Jetson only)
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)
• Issue Type( questions, new requirements, bugs)
• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)
• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)

There is no update from you for a period, assuming this is not an issue anymore.
Hence we are closing this topic. If need further support, please open a new one.
Thanks

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.