Skip to content
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

[QUESTION]: Using LibDevice - CUDA installation required on user computer? #1297

Open
AmosEgel opened this issue Nov 3, 2024 · 1 comment

Comments

@AmosEgel
Copy link

AmosEgel commented Nov 3, 2024

Question

It seems to me that using LibDevice requires the CUDA tollkit to be installed on the user computer, see here. In fact, we otherwise get errors "CUDA_PATH environment variable not found".

Is it correct that a CUDA toolkit installation is required on the user machine? Or is there a way to deploy our software (that uses ILGPU and calls LibDevice) to the user without asking them to install the CUDA toolkit?

Environment

  • ILGPU version: [e.g., 1.5.1]
  • .NET version: .NET Framework 4.7.2
  • Operating system: Windows 10
  • Hardware : NVIDIA T1000

Additional context

No response

@MoFtZ
Copy link
Collaborator

MoFtZ commented Nov 5, 2024

It is not necessary to install the CUDA SDK.

The CUDA SDK allows you to redistribute the parts you need:
https://docs.nvidia.com/cuda/eula/index.html#attachment-a

For LibDevice, if you a redistributing the CUDA files yourself, you can use the alternate LibDevice initialiser:

public Builder LibDevice(string libNvvmPath, string libDevicePath)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants