Unable to run Nsight remote monitor on linux machine using CLI

Hello everyone,

I want to run Nvidia Nsight Graphics on my linux machine to monitor a process and control it via a host(windows machine).
To get this done, I have installed Nsight graphics on both the host machine(windows) and target machine(linux).

Now, to make a connection to the target from host, it is required to run Nsight remote monitor on the remote machine.
However, in the documentation the steps mentioned to launch Nsight remote monitor are only through GUI interface. But unfortunately, I have only CLI access to my linux server.

Please help me with steps on how can I run the Nvidia Nsight remote monitor on linux using Command line Interface?

Hi anshulpandey0606,

That’s weird, Nsight Graphics is used to debug and profiler graphics apps, if you don’t have a graphics interface in Linux, how do you run your app?

Thanks
An

Hi Ayan,

Actually I am running a Deep learning application and have offloaded the inferencing part over the GPU.

I wanted to understand how the GPU side processing is happening, basically understand the scheduling of threads etc.
Based on my research, Nvidia Nsight Graphics tool seems to help with debugging and deep diving into GPU based applications.

Kindly correct me if I am wrong somewhere.
Also, I am new to GPU world so sorry for any layman queries.

Hi anshulpandey0606,

Nsight Graphics will work with some Graphics API apps, if your app using Vulkan, OGL, then you should use Nsight Graphics. If your app is using some Cuda (for deep learning), please consider choose Nsight Compute. By the way, Nsight System might also be a good candidate to you.

Thanks
An

Ohk understood. Thanks a lot for your suggestion :)

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.