Cuda and Tensorflow for SLES 15

Description

Hello,

I am stuggling to set up TensorFLow on a SLES 15.3 Server, what I mean by that is: It does not detect the GPU.

What I thought to do is:

  1. Install Nvidia Graphics Card & Drivers (you probably already have)
  2. Download & Install CUDA
  3. Download & Install cuDNN
  4. Verify by simple program

What I am not figuring out is:

where can I find the cuDNN for SUSE linux, how do I know what is compatible with what?

I do have the newest Nvidia Drivers installed the rest I am now planing to start from scratch with CUDA cuDNN. The installed CUDA Version is 12.4 is the correct one?

Environment

TensorRT Version:
GPU Type: Nvidia A2
Nvidia Driver Version: 550.54.14
CUDA Version: 12.4
CUDNN Version:
Operating System + Version: SUSE Linux Enterprise Server 15 SP3
Python Version (if applicable): 3.10.13
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag):

Hi @retusrieben ,
Please raise your issue here

Thanks

I will but this is not answering my main question about where I find out if my config is in the first place compatible.
I am talking of the CUDA version, the GPU Driver and cUDNN and so on.

I do not believe it has something to do with tensorflow but the on the NVIDIA side since PyTorch is also not working.

So I reinstalled the Drivers and also CUDA.

if I run ./deviceQuery

I get the following output:

./deviceQuery Starting…

CUDA Device Query (Runtime API) version (CUDART static linking)

cudaGetDeviceCount returned 100
→ no CUDA-capable device is detected
Result = FAIL