debug C++ project with cuda kernel using Nsight

JAMES MORGENSTERN 176 Reputation points
2025-11-29T22:06:29.1166667+00:00

Trying to configure VS 2022 community to integrate with Nvidia tools including NSight debug and compiling .cu source files. This is new install of VS 2022C and then new install of Nvidia GPU tools version v12.6. when i enter VS 2022 no signs of GPU tools, no .cu integration, no NSight available. Need help getting this straight.

Developer technologies | C++
Developer technologies | C++
A high-level, general-purpose programming language, created as an extension of the C programming language, that has object-oriented, generic, and functional features in addition to facilities for low-level memory manipulation.
{count} votes

1 answer

Sort by: Most helpful
  1. Surya Amrutha Vaishnavi Lanka (INFOSYS LIMITED) 1,115 Reputation points Microsoft External Staff
    2025-12-01T11:38:18.7033333+00:00

    Thanks for reaching out! Here are some steps you can follow

    1. Check Visual Studio Workload: Make sure you have the "Desktop development with C++" workload installed in Visual Studio. You can check this through the Visual Studio Installer.
    2. Install Nsight: Verify that Nsight is properly installed. You should be able to find it in the Visual Studio Extensions under "Manage Extensions". If it's not there, try reinstalling the NVIDIA Nsight Tools.
    3. Create a CUDA Project: After verifying the installation, try creating a CUDA project through File > New > Project > and look for CUDA templates. This will help you see if CUDA-specific options are available.
    4. Update Environment Variables: Ensure that the CUDA and Nsight binaries are included in the system PATH environment variable. This could be crucial for integration.
    5. Registry Keys and Directories: Check that the CUDA installation directory exists (default is usually C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6) and that the registry keys for CUDA are correctly set. You can check the registry entries under HKEY_LOCAL_MACHINE\SOFTWARE\NVIDIA Corporation\CUDA.
    6. Visual Studio Extensions: Sometimes, disabling all extensions temporarily and only enabling Nsight can resolve conflicts.

    If you’ve tried these steps and are still facing issues, it would help to know:

    • Can you confirm that the CUDA installation itself is functioning correctly? (e.g., running a simple CUDA sample project in the command line)
    • Have you tried running Visual Studio as an Administrator?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.