createInferRuntime in dll get stuck forever

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Description

I have the following two lines of code in my dll, but found that when running to the second line, it blocks there forever and never returns，how can i fix this?

static std::unique_ptr<Logger> s_logger = std::make_unique<Logger>();
static std::unique_ptr<nvinfer1::IRuntime> s_runtime = std::unique_ptr<nvinfer1::IRuntime>(nvinfer1::createInferRuntime(*s_logger));

Environment

TensorRT Version: 8.4.3.1
GPU Type: dGPU
Nvidia Driver Version: 531
CUDA Version: 11.6
CUDNN Version: 8.4
Operating System + Version: windows10
Python Version (if applicable):
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag):

Relevant Files

Please attach or include links to any models, data, files, or scripts necessary to reproduce your issue. (Github repo, Google Drive, Dropbox, etc.)

Steps To Reproduce

Please include:

  • Exact steps/commands to build your repro
  • Exact steps/commands to run your repro
  • Full traceback of errors encountered
2

Hi,
Please check the below links, as they might answer your concerns.

Thanks!

4

Sorry, I’m afraid the above documentation isn’t much help.
I’ve submitted an attached document that reproduces the problem, please take a look at it
Project4.zip (21.0 MB)

5

Heeeeelp, anybody any idea?

6

Could you please pass a simple Logger object here and check if it works.
Example: Developer Guide :: NVIDIA Deep Learning TensorRT Documentation

7

Thank you for your reply.

I changed the Logger to the simplest form as you suggested, but the result is the same, it keeps blocking the creation of the Runtime!

Here are some experiments I did.

  1. change tensorrt to version 8.6, it doesn’t work.
  2. move the runtime initialization process from a dll library to an exe executable, it works.
  3. port the above code to ubuntu, it works.

Any other suggestions?