[libcu++] Dynamically load CUDA library instead of using the runtime #6899
+30
−10
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
|
Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually. Contributors can view more details about this message here. |
Contributor
Author
|
/ok to test 91bea5b |
pciolkosz
force-pushed
the
dlopen_libcuda_so
branch
from
91bea5b to
95edf73
Compare
pciolkosz
force-pushed
the
dlopen_libcuda_so
branch
from
95edf73 to
319d34b
Compare
Contributor
Author
|
/ok to test 319d34b |
This comment has been minimized.
This comment has been minimized.
Contributor
Author
|
/ok to test 35dd82a |
Contributor
😬 CI Workflow Results🟥 Finished in 1h 48m: Pass: 93%/91 | Total: 1d 04h | Max: 1h 46m | Hits: 96%/199476See results here. |
miscco
requested changes
Dec 8, 2025
Comment on lines
+56
to
+58
| # if _CCCL_OS(WINDOWS) | ||
| HMODULE m_cudaDriverLibrary = LoadLibraryEx("nvcuda.dll", nullptr, LOAD_LIBRARY_SEARCH_SYSTEM32); | ||
| if (m_cudaDriverLibrary == nullptr) |
Contributor
There was a problem hiding this comment.
Critical: This is loading every single time the function is called, should this be a function local static?
| # else | ||
| const char* m_cudaDriverLibraryName = "libcuda.so.1"; | ||
| # endif | ||
| void* m_cudaDriverLibrary = dlopen(m_cudaDriverLibraryName, RTLD_NOW); |
Contributor
There was a problem hiding this comment.
ditto should this be function local static?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
We can't support minor version compatibility in 12.X releases without breaking dynamic runtime use case described here: #5970. Its because in older 12.X releases there is no
cudaGetDriverEntryPointByVersionand the non versioned one can't support MVC.This PR switches to instead dynamically load the CUDA library and fetch
cuGetProcAddressfrom it, instead of usingcudaGetDriverEntryPoint