Hello, I Have 4 Python Environments On One Server With 4 Tensorflow Versions. In Each Python Environment The Correct Cuda Version Is Used. But When I Use Allegro The Error Comes: Could Not Load Dynamic Library 'Libcublas.So.10'; Dlerror: Libcublas.So.10:

Answered

hello,
i have 4 python environments on one server with 4 tensorflow versions. In each python environment the correct cuda version is used.
But when I use allegro the error comes:
Could not load dynamic library 'libcublas.so.10'; dlerror: libcublas.so.10: cannot open shared object file: No such file or directory
There is a problem with the path variable. Can I change this in Allgero?
I can't use Docker because I need 4 different Tensorflow versions and my company is not allowed to use conda.

  				
Posted 
	3 years ago

					More
				  		
  Report
		
					UnsightlySeagull42
				
					0
					 × 1

Votes Newest

Answers 3

Is there a way how to fix it?

  				
Posted 
	3 years ago

					More
				  		
  Report
		
					UnsightlySeagull42
				
					0
					 × 1

Hi UnsightlySeagull42 , didnt really get your setup, you have more than one cuda on your system?

I can’t use Docker because I need 4 different Tensorflow versions and my company is not allowed to use conda.

You can use Docker without conda

  				
Posted 
	3 years ago

					More
				  		
  Report
		
					TimelyPenguin76
				
					0
					 Administrator

Thank you for the answer.
I have 2 different cuda versions.
I need tensorflow 2.2, 2.3, 2.4, 2.5.
For tensorflow 2.2 i need cuda 10.1
But for tensorflow 2.4 i need for example cuda 11.0
https://www.tensorflow.org/install/source#gpu .
For docker I use for example: --docker nvidia/cuda:10.1-cudnn7-runtime-ubuntu18.04
Then tensorflow 2.4 no longer works because tensorflow 2.4 requires cuda 11 and not cuda 10.1
Does anyone have any idea?
I can also pass 2 different docker images?

  				
Posted 
	3 years ago

					More
				  		
  Report
		
					UnsightlySeagull42
				
					0
					 × 1

Write your answer

1K Views

3 Answers

3 years ago

one year ago