Skip to content

CudaProvider jumped completely when Tensorrt and Openvino provided #25347

Open
@aipatr

Description

@aipatr

Describe the issue

Dear all,
I compiled the onnxruntime starting from the main branch on Windows 11 using the Visual Studio Compiler. I used:

  • CUDA libraries installed (specifically CUDA 12.6)
  • cuDNN library ( cuDNN 9.5)
  • TensorRT library (version 10.11.0.33)
  • OpenVINO library (version 2025.2.0 or compatible)

I created the wheel with the following command:
.\build.bat --config Release --build_shared_lib --parallel --use_tensorrt --use_openvino --tensorrt_home="C:\codebase\onnxruntime_compiled\TensorRT-10.11.0.33" --cuda_home="C:\codebase\piur-tus-4\PiurCore\third_party\CUDA\v12.6" --cudnn_home="C:\codebase\piur-tus-4\PiurCore\third_party\CUDA\v12.6" --skip_tests --compile_no_warning_as_error --enable_cuda_minimal_build --use_vcpkg --enable_pybind --build_wheel

Tensorrt works
OpenVino works
CudaProvider is available but ignored when selected and CPU is used instead

To reproduce

Compiled the onnxruntime starting from the main branch on Windows 11 using the Visual Studio Compiler. using:

  • CUDA libraries installed (specifically CUDA 12.6)
  • cuDNN library ( cuDNN 9.5)
  • TensorRT library (version 10.11.0.33)
  • OpenVINO library (version 2025.2.0 or compatible)

Create the wheel with the following command:
.\build.bat --config Release --build_shared_lib --parallel --use_tensorrt --use_openvino --tensorrt_home="C:\codebase\onnxruntime_compiled\TensorRT-10.11.0.33" --cuda_home="C:\codebase\piur-tus-4\PiurCore\third_party\CUDA\v12.6" --cudnn_home="C:\codebase\piur-tus-4\PiurCore\third_party\CUDA\v12.6" --skip_tests --compile_no_warning_as_error --enable_cuda_minimal_build --use_vcpkg --enable_pybind --build_wheel

Tensorrt works
OpenVino works
CudaProvider is available but ignored when selected and CPU is used instead

Urgency

No response

Platform

Windows

OS Version

11

ONNX Runtime Installation

Built from Source

ONNX Runtime Version or Commit ID

1.23

ONNX Runtime API

Python

Architecture

X64

Execution Provider

CUDA

Execution Provider Library Version

No response

Metadata

Metadata

Assignees

No one assigned

    Labels

    ep:OpenVINOissues related to OpenVINO execution providerep:TensorRTissues related to TensorRT execution provider

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions