Fix CUDA minimal build requiring CUDNN_HOME#27308
Merged
tianleiwu merged 1 commit intomicrosoft:mainfrom Feb 12, 2026
Merged
Conversation
When building with onnxruntime_CUDA_MINIMAL=ON, cuDNN is not needed since minimal CUDA builds only use basic CUDA runtime (e.g., for TensorRT-only builds). Skip cuDNN.cmake inclusion and CUDNN_INCLUDE_DIR references when CUDA_MINIMAL is enabled. Fixes microsoft#24361
tianleiwu
approved these changes
Feb 11, 2026
Contributor
|
/azp run Linux QNN CI Pipeline, Win_TRT_Minimal_CUDA_Test_CI, Windows ARM64 QNN CI Pipeline, |
|
Azure Pipelines successfully started running 4 pipeline(s). |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When building with
onnxruntime_CUDA_MINIMAL=ON, cuDNN is not needed since minimal CUDAbuilds only use basic CUDA runtime (e.g., for TensorRT-only builds). Skip
cuDNN.cmakeinclusion and
CUDNN_INCLUDE_DIRreferences whenCUDA_MINIMALis enabled.Fixes #24361
Description
Make cuDNN dependency conditional on
NOT onnxruntime_CUDA_MINIMAL:cmake/external/onnxruntime_external_deps.cmake: Skipinclude(cuDNN)cmake/onnxruntime_python.cmake: SkipCUDNN_INCLUDE_DIRin include pathscmake/onnxruntime_unittests.cmake: SkipCUDNN_INCLUDE_DIRin two placesMotivation and Context
When building with
onnxruntime_CUDA_MINIMAL=ON(e.g., TensorRT-only builds), the buildfails with:
CMake Error at external/cuDNN.cmake:3 (find_path):
Could not find CUDNN_INCLUDE_DIR using the following files: cudnn.h
cuDNN is not needed for minimal CUDA builds since TensorRT has its own optimized kernels.
This fix allows building without setting
CUDNN_HOME.