VisionWorks on Nvidia Nsight, how to do?

Hello guys,
I’ve been searching for a solution from a while now but I had no luck. Is there anyone who have managed to get VisionWorks working on the Nsight IDE? Is writing code directly on the Jetson TK1 the only solution!?

Thank you in advance.

Hello Marcello-barile,

Did you try to follow the guidance here:

The key thing is after launching Nsight, you just need to use the option “File/Import/C/C++/Existing Code as Makefile Project” and then browse to your local VisionWorks sample code, select suitable CUDA Toolkit (7.5 or 8.0), then click Finish button.

You can install the VisionWorks sample code by going to /usr/share/visionworks/sources and launch below command:

$ ./ ~/

to have the VisionWorks samples to be installed to your home folder.

I am trying to do this for Nano.
I tried the above and imported the file system for example ‘hough transform’. It still needs to include all the lib, data and other folders in the same project. So just importing the relevant project is not sufficient. Unlike CUDA demos, there is not ‘common’ folder in the VisionWorks samples.
Looked up this one :

Still, no conclusive success in building and running through SSH in cross-compile mode. (is this possible for VisionWorks samples?)
Requesting some directives needed on how to build the visionwork projects from within the Nsight Eclipse?
Please advise.