CUDA initialization failure with error 35 (TX2)

Autonomous Machines Jetson & Embedded Systems Jetson TX2
,
#1

Hi,
There are many TX2s used in the project. If each TX2 uses SDKmanger to refresh the system and cuda, it will take a lot of time.

So try to refresh a TX2 system(Jetpack4.4.0) by using SDKmanger, and then get the dynamic libraries of CUDA and TensorRT. Other TX2 directly use the dynamic libraries of CUDA and TensorRT, but errors occur: "Cuda initialization failure with error 35. Please check cuda installation: http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html."

Cuda dynamic library obtained by SDKmanger(Ubuntu18.04.04), not be used in TX2 native system(Ubuntu18.04.02)?
Are there other ways to deploy TX2 system software environment in batches?
Thanks.

#3

Hi,

The error indicates a compatibility issue.

You will need to use the package downloaded from the SDK manager.
The package from the website is usually for desktop (x86) users.

Thanks

#4

Hi,

I did not use this website to download the software, but reported an error to show this website, which has nothing to do with this topic.

the question is that, try to refresh a TX2 system(Jetpack4.4.0) by using SDKmanager, and then get the dynamic libraries of CUDA and TensorRT. Other TX2(no using SDKmanager) directly use the dynamic libraries of CUDA and TensorRT, but errors occur! Is this transplantation method okay?

Or is there any other way to transplant the TX2 system environment? Instead of using every tx2 system by using SDKmanager.
Thanks

#5

oh,
Can use L4T Jetson Driver Package to copy TX2 system: https://developer.nvidia.com/zh-cn/embedded/downloads