How to run DeepStream SDK Samples in NSight ?

We brought brand new NVIDIA Titan Xp for prototyping a Deep Stream Object Detection Module. We played around with all sample codes which comes in DeepStream SDK. Now we are customizing it add some of our custom Pipelines. We used GDB and were able to debug it. Since command line tool is not handy for the entire team, we thought of using NSight for better productivity.

So we have below directory structure -

user1@ubox:~/SDK/deepstream$ tree -L 2
.
├── doc
│   ├── DeepStream_User_Guide.pdf
│   └── Release_notes.txt
├── include
│   ├── deepStream.h
│   └── module.h
├── lib
│   └── libdeepstream.so
├── LicenseAgreement.pdf
├── makes
│   └── defines.inc
└── samples
    ├── bin
    ├── common
    ├── data
    ├── decPerf
    ├── Makefile
    ├── nvDecInfer_classification
    ├── nvDecInfer_detection
    └── nvSmartDecInfer_detection

It would great if you help -

  1. How to import entire Samples into NSight ?
  2. What are the run configurations we need to make in order to build and run in NSight ? we used run like this
    @ubox:~/SDK/deepstream/samples$ make
    @ubox:~/SDK/deepstream/samples/nvDecInfer_detection$ ./run.sh
    Success in running it!
    

We are Visual Studio Developer, NSight very new to all of us(never used before). So it would be really great if you help in detailed steps.

Thank you
Ajay

Hi,

Here is our Nsight tutorial video:
https://www.youtube.com/watch?v=FAgq8VSWJ34

In general, libraries information can be added via:
Build -> Settings -> Tool Settings
NVCC Compiler -> Includes -> Include paths -> Add include path
NVCC Linker -> Libraries -> Libraries -> Add required libraries
NVCC Linker -> Libraries -> Library search path -> Add library path

Here is another tutorial to cross-compile an OpenCV project on Jetson for your reference:
https://devtalk.nvidia.com/default/topic/1010145/jetson-tx1/how-to-use-opencv3-1-with-nvidia-nsight-/post/5157413/#5157413

Thanks.