TensorRT on Windows 10 with CUDA 11.5 and cuDNN 8.3

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

Description

Hi everyone, I’m trying to install TensorRT on my Windows machine, in the Installation Guide it says:

Ensure that you have the following dependencies installed.
- CUDA 10.2, 11.0 update 1, 11.1 update 1, 11.2 update 2, 11.3 update 1, 11.4 update 3 or 11.5
- cuDNN 8.2.1

Given that I have cuDNN 8.3 installed, do I have to downgrade it to cuDNN 8.2.1 or will it work anyways? Moreover, supposing that I will serialize an optimized torch script module through TensorRT, will I be able to use it for inference even on a machine with a different version of CUDA (e.g. 10.2) ?

Environment

TensorRT Version: 8.2.1.8
GPU Type: NVIDIA GeForce RTX 3080
Nvidia Driver Version: 496.13
CUDA Version: 11.5
CUDNN Version: 8.3.0
Operating System + Version: Windows 10
Python Version (if applicable): 3.10
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag):

Relevant Files

None

2

Hi,
Please refer to the installation steps from the below link if in case you are missing on anything

Also, we suggest you to use TRT NGC containers to avoid any system dependency related issues.

Thanks!