Nvidia Nsight Host Machine Requirements

Accelerated Computing CUDA CUDA Setup and Installation
1

Hi All.

I just had a quick question about remote debugging a CUDA project using Nvidia Parralell Nsight.

Does Nsight require a Nvidia device installed on the host machine in order to operate?

I have a desktop with a GTX650 installed (my target machine) but the only other (host) machine I currently have access to has an AMD device installed on it. Does this mean I cannot use this machine to debug my project?

If the worst comes to the worst I can always setup a dual GPU debugging enviornment with the onboard output from my target machine but I would rather not go through that if I don’t have to.

Any response is greatly appreciated. Thanks.

2

Con07,

The requirements for local and remote debugging are documented in the Nsight Visual Studio Edition User Manual in the section System Requirements for NVIDIA Nsight Software url[/url].

CUDA remote debugging does not require the host to have an NVIDIA GPU.

Nsight Visual Studio Edition 2.2 introduced single GPU CUDA debugging.

I would recommend that you upgrade to Nsight 3.0 RC1 available for download through the Nsight Visual Studio Edition Registered Developer Program.