Crash bug of running the deepstream and performance optimization issues on jetson xavier NX

Accelerated Computing Intelligent Video Analytics DeepStream SDK
,
1

• Hardware Platform (Jetson / GPU)：Jetson Xavier NX
• DeepStream Version：6.3
• JetPack Version (valid for Jetson only)：5.1.2
• TensorRT Version：8.5.2-1
• Issue Type( questions, new requirements, bugs)：questions and bugs
• Bug：
I want to deploy custom YoloV8 model in Deepstream app, so I followed this tutorial to install and test.(The only different is that I didn’t simplify the ONNX model)

DeepStream-Yolo/docs/YOLOv8.md at master · marcoslucianops/DeepStream-Yolo (github.com)

I can run the program successfully once. However, if I want to run this command line

deepstream-app -c deepstream_app_config.txt

again. My jetson Xavier NX will crash, all processes will be unresponsive, and I can only move the mouse cursor, mouse clicks and keyboard input are unresponsive.
I tried many times. The only way out is to power down and restart my NX.
I observed the power GUI and found that my GPU would be full while running, but there were no noticeable anomalies with the CPU or memory, and the GPU would return to normal after ending the program, so I couldn’t find a reason for the crash at all.

Monitor while running the program.

cefeb99b6eb63cc27b40da2d9a14d87
cefeb99b6eb63cc27b40da2d9a14d871920×1080 145 KB

Monitor after ending the program.

eca769ed8ce8cd0038efea02ba8be98
eca769ed8ce8cd0038efea02ba8be981920×1080 138 KB

• Question：

  1. Is there now an official tutorial for deploying yolov8 on deepstream, I’m sorry I can’t seem to find one, I’m wondering if there is an official tutorial to fix my bug.
    1. How can I Adjusting my parameters to improve my frame rate results? I use default parameter to run it, now its frame rate is about 20.
    1. If I run the program, it will notice me that “system throttled due to over-current.”,and the logo in the upper right corner will change to an exclamation point. It’s shown below. What does that mean?
  4. d1fb0c0a26076026945149f84d8f384
    d1fb0c0a26076026945149f84d8f3841514×1025 69 KB
3

noticing your GPU would be full while running, When the GPU usage is about 100%, it can cause your screen to lag because your GPU is fully occupied with computing tasks. As a result, it may not have enough resources to render the graphical user interface (GUI) smoothly.

  1. please refer to YOLO FAQ.
  2. please refer to topic for performance analysis.
  3. please refer to this topic.
4

Thank you. I have solved the crash problem following the third topic, but it still noticed me that “system throttled due to over-current.”
Can you offer me a better solution for it? It seems not be solved well in that topic; many users are still troubled by it.
Thank you again!

5

please refer to topic 173571. and “system throttled due to over-current” issue would be outside of DeepStream, please open a new topic in forum xavier-nx section. Thank you!