Why the Deepstream 5.0.1 SDK runs 40 cameras and the FPS is very low, but the CPU and GPU occupancy is very low？

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Please provide complete information as applicable to your setup.

**• Hardware Platform (Jetson / GPU)Tesla T4
**• DeepStream Version 5.0.1
• JetPack Version (valid for Jetson only)
**• TensorRT Version 7.0.0
**• NVIDIA GPU Driver Version (valid for GPU only)10.2

below are some of my images and config files
CPU:

cpu
cpu1666×508 81.1 KB

cpu2
cpu21751×194 107 KB

cpu3
cpu31467×205 281 KB

GPU:
gpu
gpu1349×545 76.6 KB

Run Result:
result
result1803×180 42.9 KB

Config File:
deepstream_app_config_yoloV5_Helmet.txt (35.0 KB)

3

Can you use “nvidia-smi dmon” to monitor the GPU and CPU loading?

4

encoder
encoder1349×313 99.9 KB

thanks.Is this because the encoder of the graphics card is already fully loaded? Has nvidia officially tested the maximum number of rtsp streams that the Tesla T4 can encode at the same time?

5

The HW encoder is fully occupied by the 40 streams encoding for RTSP streaming.

6

I know,how many rtsp streams can be HW encode in Tesla T4?

7

Depends on the resolution and encoding parameters. You can try to test.

8

thanks