Windows OS

How to Check If TensorFlow Is Using NVIDIA CuDNN on Windows 10/11

NVIDIA cuDNN or NVIDIA CUDA Deep Neural Network is a library of common functions that is used to accelerate the AI/ML training using NVIDIA GPUs.

In this article, we will show you how to check whether TensorFlow can access the NVIDIA cuDNN libraries. This helps you to verify whether the NVIDIA cuDNN library is successfully installed on your Windows 10/11 system.

If you read this article to install NVIDIA cuDNN on your Windows 10/11 system, the NVIDIA cuDNN DLL files should be in the C:\Program Files\NVIDIA\CUDNN\CUDA12_8.x\bin (for cuDNN for CUDA 12) or C:\Program Files\NVIDIA\CUDNN\CUDA11_8.x\bin (for cuDNN for CUDA 11) folder path. You will find the “cudnn64_8.dll” DLL file in this path.

To verify whether TensorFlow can access the NVIDIA cuDNN library files, you can check whether Python can find the NVIDIA cuDNN “cudnn64_8.dll” DLL file. If Python can find this DLL file, it will also be able to find the other NVIDIA cuDNN DLL files that you can see in the following screenshot:

First, open a Terminal app on Windows 10/11 and run the following command to access the Python CLI interpreter:

$ python

Then, import the “ctypes” Python module and check whether the NVIDIA cuDNN “cudnn64_8.dll” DLL file is accessible with the following lines of codes:

$ import ctypes

$ ctypes.WinDLL("cudnn64_8.dll")

If Python can access the NVIDIA cuDNN library DLL files, it prints the handle ID and the memory location where the “cudnn64_8.dll” DLL file is loaded. Since Python can access the NVIDIA cuDNN library DLL files, TensorFlow can also access it. It means that TensorFlow can also use the NVIDIA cuDNN library functions on your Windows 10/11 system.

A black background with white text Description automatically generated


In this article, we showed you how to verify whether TensorFlow can use the NVIDIA cuDNN library on the Windows 10/11 operating systems.

About the author

Shahriar Shovon

Freelancer & Linux System Administrator. Also loves Web API development with Node.js and JavaScript. I was born in Bangladesh. I am currently studying Electronics and Communication Engineering at Khulna University of Engineering & Technology (KUET), one of the demanding public engineering universities of Bangladesh.