Hi amycao,
Thank you for your reply
I don’t think it’s problem about DISPLAY
xrandr
Screen 0: minimum 8 x 8, current 1920 x 1200, maximum 32767 x 32767
HDMI-0 disconnected (normal left inverted right x axis y axis)
DP-0 disconnected (normal left inverted right x axis y axis)
DP-1 disconnected (normal left inverted right x axis y axis)
DP-2 connected primary 1920x1200+0+0 (normal left inverted right x axis y axis) 518mm x 324mm
1920x1200 59.95*+
1920x1080 60.00
1680x1050 59.95
1600x1200 60.00
1280x1024 60.02
1280x960 60.00
1024x768 60.00
800x600 60.32
640x480 59.94
And gst-launh-c1.0 videotestsrc ! nveglglessink could show the display
But I still failed on deepstream-app -c source1_usb_dec_infer_resnet_int8.txt
NOTE: To expand a source in the 2D tiled display and view object details, left-click on the source.
To go back to the tiled display, right-click anywhere on the window.
Deepstream requires nvidia GPU for output to display, if you are using Tesla card, please noted it’s compute card, but you can follow this to setup virtual display, Deepstream/FAQ - eLinux.org
5A, if nvidia GPU with display port, please check your display driver environments;
you also can use sink type 4 RTSPStreaming for visulization on host by vlc.
Hi amycao,
I see,my card is 1080ti. I always want to develop on desktop first and then transfer to jetson. I think i could use rtsp,it‘s ok. Thank you for answer
Hi amycao, thanks for your follow-up. I have run the test again following your new instructions (with sink2 RTSPStreaming), and I got the below output, now I need to know how to visualize the streamed output pointed at *** DeepStream: Launched RTSP Streaming at rtsp://localhost:8080/ds-test ***
Hi Amycao thanks for the reply, I decided to save the out video to a file disabling sink0 and enabling sink1, and it is working fine.
I have a server with 6xT4’s. When I run deepstream-app on just 1 GPU the performance is around 39 FPS; however, I decided to run the application in parallel on each GPU and the performance dropped to 16 FPS for each GPU. Do you have an idea about the root cause for the dropping of performance?
Configuration to save the output video to a file:
[sink0]
enable=0 #Type - 1=FakeSink 2=EglSink 3=File
type=2
sync=1
source-id=0
gpu-id=0
nvbuf-memory-type=0
Configuration to run the application on each GPU:
-Start the container image on 6 different shell windows
-Set a different gpu id for each container image with the flag gpu-id=<GPU ID #>
-Run the application parallely
from your config, sink1 used, it’s for File type, which will save the output onto your disk, it will not take GPU, through docker to set gpu id? or can you paste the whole config?
Hello @Amycao,
I have a very similar problem related to this post. My platform details are - • Hardware Platform GPU (Tesla K20Xm) • DeepStream Version - 5.0 • NVIDIA GPU Driver Version-440.33.01
I am trying to run a test application with a name deepstream-nvdsanalytics-test in /opt/nvidia/deepstream/deepstream-5.0/source/apps/sample-apps/deepstream-nvdsanalytics-test. It uses dsanalytics plugin for processing line-crossing, direction detection, overcrowding etc.
I followed this post and was able to run the test by doing following changes in deepstream_nvdsanalytics_test.cpp-
I got all information (overcrowding, line-crossing, direction) printed in the terminal and also got video file. But nothing got written in that video file “capture.mp4”. The file was completely empty with a size of few KB. I am not able to understand what is going wrong.
Please help, and let me know if any more information is required.
Thanks in advance.
Hello @sanchezvr7 please let me know if you are able to implement the solution suggested by @Amycao.
I went through the code and found it difficult to understand. I am still trying and if I understand the implementation I will post.
Thank you for the response.