Skip to content

[Build] Fix cuda link target of cumem_allocator in CPU env #12863

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.

Sign up for GitHub

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

Merged
merged 2 commits into from
Feb 11, 2025
Merged

[Build] Fix cuda link target of cumem_allocator in CPU env #12863

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3b09f0c
Fix cuda link target of cumem_allocator
guoyuhong Feb 6, 2025
c3c60ab
Merge branch 'main' into fix_cumem_allocator_cmake
tlrmchlsmth Feb 10, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -192,7 +192,7 @@ set_gencode_flags_for_srcs(
if(VLLM_GPU_LANG STREQUAL "CUDA")
message(STATUS "Enabling cumem allocator extension.")
# link against cuda driver library
list(APPEND CUMEM_LIBS cuda)
list(APPEND CUMEM_LIBS CUDA::cuda_driver)
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This seems reasonable to me -- @youkaichao WDYT?

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actually can we just delete this line completely? We have this link line in define_gpu_extension_target already:

vllm/cmake/utils.cmake

Lines 437 to 443 in 2ae8890

# Don't use `TORCH_LIBRARIES` for CUDA since it pulls in a bunch of
# dependencies that are not necessary and may not be installed.
if (GPU_LANGUAGE STREQUAL "CUDA")
target_link_libraries(${GPU_MOD_NAME} PRIVATE CUDA::cudart CUDA::cuda_driver)
else()
target_link_libraries(${GPU_MOD_NAME} PRIVATE ${TORCH_LIBRARIES})
endif()

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Per my analysis, only when GPU_LANGUAGE STREQUAL "CUDA" the CUDA::cudart lib is linked in define_gpu_extension_target. However, the GPU_LANGUAGE for cumem_allocator is CXX. I think that is why extra libcuda is added to CUMEM_LIBS.

define_gpu_extension_target(
cumem_allocator
DESTINATION vllm
Expand Down