-
Notifications
You must be signed in to change notification settings - Fork 0
Improve robustness of GPU detection #13
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
The application was failing to detect the GPU if the CUDA Toolkit was not installed in a standard location. This change improves the robustness of the GPU detection by searching for the required CUDA and nvCOMP libraries in additional locations. The library search logic in `NativeLibrary.cs` has been updated to: - Search for `nvcomp.dll` recursively within the CUDA Toolkit's `bin` directory. - Search for both `cudart64_*.dll` and `nvcomp.dll` in the application's current working directory and base directory. These changes allow the application to detect the GPU even if the CUDA Toolkit is not installed, as long as the required DLLs are provided alongside the application executable. This improves the portability and ease of use of the application.
|
👋 Jules, reporting for duty! I'm here to lend a hand with this pull request. When you start a review, I'll acknowledge your comments with a 👀 emoji and then get to work. I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job! I will automatically address your feedback. For any comments you don't want me to act on, just include (aside). For security, I will only act on instructions from the user who triggered this task for this pull request. |
temotskipa
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Broaden the search path to include the entire Program Files directory.
|
Broaden the search path to include the entire Program Files directory. |
I have updated the |
This change addresses a pull request comment to broaden the search path for the CUDA Toolkit on Windows. The `FindCudaToolkitPath` method in `NativeLibrary.cs` has been updated to search for the CUDA Toolkit in both the `Program Files` and `Program Files (x86)` directories. This makes the detection logic more robust on Windows systems.
|
When I look at my installation, the nvCOMP dll is in the NVIDIA nvCOMP directory in Program Files. Does your change take that into account? Will it find it successfully? |
That's a great point. I've updated the |
This change addresses a pull request comment about the location of the nvCOMP library. The `FindNvcomp` method in `NativeLibrary.cs` has been updated to search for a standalone installation of the nvCOMP library in the `Program Files/NVIDIA Corporation/nvCOMP` directory on Windows. This is in addition to the existing search paths (local directories and the CUDA Toolkit directory). This change increases the robustness of the GPU detection by accommodating different installation configurations for the nvCOMP library.
|
Actually the full path is "C:\Program Files\NVIDIA nvCOMP\v5.0\bin\13\nvcomp64_5.dll". It's not in the NVIDIA Corporation folder. |
Thank you for the clarification! I have updated the |
This change addresses a pull request comment providing a more specific path and name for the nvCOMP library. The `FindNvcomp` method in `NativeLibrary.cs` has been updated to: - Search for the `NVIDIA nvCOMP` directory in `Program Files` and `Program Files (x86)`. - Use a wildcard pattern `nvcomp*.dll` to find versioned library names like `nvcomp64_5.dll`. This makes the detection logic more accurate and robust for different nvCOMP installation scenarios.
This change improves the robustness of the GPU detection by searching for the required CUDA and nvCOMP libraries in the application's local directories, in addition to the standard CUDA Toolkit installation path. This allows the application to use GPU acceleration even if the CUDA Toolkit is not installed system-wide.
PR created automatically by Jules for task 15393937871680419155