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

%CUDA_PATH% activation on Windows #232

Closed
jakirkham opened this issue Nov 20, 2023 · 6 comments
Closed

%CUDA_PATH% activation on Windows #232

jakirkham opened this issue Nov 20, 2023 · 6 comments

Comments

@jakirkham
Copy link
Member

Currently on Windows we set %CUDA_PATH% to %CONDA_PREFIX%

set "CUDA_PATH=%CONDA_PREFIX%"

However it appears when users run with this this, they get CUDA_PATH detection warnings. Though they seem to be able to use CuPy without issues

Since the headers and includes from the CTK typically are installed to %CONDA_PREFIX%\Library, wonder if we should update this to point to that location

-set "CUDA_PATH=%CONDA_PREFIX%"
+set "CUDA_PATH=%CONDA_PREFIX%\Library"

Though looking at this logic in CuPy, am not sure if we should point to %CONDA_PREFIX%\Library\bin instead

Thoughts?

@jakirkham
Copy link
Member Author

It's worth noting there is also some discussion about updating or dropping the activation script ( #206 ), which might also affect how this is handled

@leofang
Copy link
Member

leofang commented Nov 20, 2023

Let us move the discussion to #206, I think it's the same issue.

@leofang
Copy link
Member

leofang commented Nov 20, 2023

See #206 (comment).

@jakirkham
Copy link
Member Author

Ok I read that as Linux whereas the issue here is Windows

The targets point specifically doesn't apply to Windows

On Windows, we have a more similar CTK structure between CUDA 11 / 12

@leofang
Copy link
Member

leofang commented Nov 20, 2023

Ok I read that as Linux whereas the issue here is Windows

The targets point specifically doesn't apply to Windows

On Windows, we have a more similar CTK structure between CUDA 11 / 12

To CuPy the same applies to Linux as well. All headers and shared libraries can be found in the targets directory, enough to ensure a fully functional CuPy runtime.

@jakirkham
Copy link
Member Author

Ok let's close then

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

No branches or pull requests

2 participants