Very less throughput of deepstream sample applications


I am using Jetson nano 4GB Dev Kit with following configuration.

•Hardware Platform (Jetson / GPU) : Jetson Nano 4GB Dev Kit
•DeepStream Version : 5.1
•JetPack Version (valid for Jetson only) : 4.5.1
•TensorRT Version : 7.1.3
•CUDA Version: 10.2
•cuDNN Version: 8.0

I am trying to run the deepstream sample application using trafficcamnet configuration file with sample streams of deepstream SDK or our custom stored video (640x480). I am using the command show below:

$ deepstream-app -c deepstream_app_source1_trafficcamnet.txt 

Initially application start working fine But after few seconds the system gives system error message as below:

system throttled due to overcurrent 

The application starts dropping the buffers while processing and some times it works fine for few seconds. The application gives following warning:

**PERF:  17.87 (16.78)	
WARNING from sink_sub_bin_sink1: A lot of buffers are being dropped.
Debug info: gstbasesink.c(2902): gst_base_sink_is_too_late (): /GstPipeline:pipeline/GstBin:processing_bin_0/GstBin:sink_bin/GstBin:sink_sub_bin1/GstEglGlesSink:sink_sub_bin_sink1:
There may be a timestamping problem, or this computer is too slow.

Because of this issue system throughput is very less not the real time processing is happening. I am having jetson nano devkit with only passive heat sink without fan mounted on the top.

I tested following test sample application too:

  1. deepstream-test2
  2. deepstream-nvdsanalytics-test

I am getting the same throughput with above application too.

Please help us to fix this problem.

Vikas Dwivedi

Seems system issue. Please refer below topic:

Hello kesong,

I found two things to be tried:

  • Increase the current limit

  • Reduce the CPU/GPU frequency to lower down the power consumption.

Please tell me the command to increase the current limit in jetson nano.

I have gone through the Power Estimator Tool to create the customized power mode but unfortunately it is not available for Jetson Nano.

Please let me know the process to make custom power mode and how we can reduce the CPU/GPU frequency of jetson Nano.

Thanks and Regards,
Vikas Dwivedi

Hi @vikas.wfh ,

Can you look into this issue and see if your configuration is correct and you are able to get adequate Voltage and Power.


May I know what is your nvpmodel setting now?

Hello Mr. Giddwani,

Thank you for your suggestions.

Problem got resolved by replacing 5V adaptor with 5.5V 10A.

Thanks and Regards,
Vikas Dwivedi