# Cuda 12.1, TensorFlow on Linux 20.04, does not see GPU Nvidia 4090 RTX

**URL:** <https://forums.developer.nvidia.com/t/cuda-12-1-tensorflow-on-linux-20-04-does-not-see-gpu-nvidia-4090-rtx/252514>\
**Category:** CUDA Setup and Installation\
**Created:** [May 7, 2023, 12:43am UTC](https://forums.developer.nvidia.com/t/cuda-12-1-tensorflow-on-linux-20-04-does-not-see-gpu-nvidia-4090-rtx/252514 "2023-05-07T00:43:24Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![michael370](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@michael370](https://forums.developer.nvidia.com/u/michael370)\
**Post date:** [May 7, 2023, 12:43am UTC](https://forums.developer.nvidia.com/t/cuda-12-1-tensorflow-on-linux-20-04-does-not-see-gpu-nvidia-4090-rtx/252514/1 "2023-05-07T00:43:24Z")

</div>

python3 -c “import tensorflow as tf; print(tf.config.list\_physical\_devices(‘GPU’))”  
tensorflow/tsl/cuda/cudart\_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.

Yes, I see on the tensorflow website, that only tested with lower CUDA, Using Python 3.10, PyTOrch cuda\_is\_available,

nvcc --version  
nvcc: NVIDIA (R) Cuda compiler driver  
Copyright (c) 2005-2019 NVIDIA Corporation  
Built on Sun\_Jul\_28\_19:07:16\_PDT\_2019  
Cuda compilation tools, release 10.1, V10.1.243

nvidia-smi  
Sat May 6 17:42:47 2023  
±--------------------------------------------------------------------------------------+  
| NVIDIA-SMI 530.41.03 Driver Version: 530.41.03 CUDA Version: 12.1 |  
|-----------------------------------------±---------------------±---------------------+  
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |  
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |  
| | | MIG M. |  
|=========================================+======================+======================|  
| 0 NVIDIA GeForce RTX 4090 Off| 00000000:41:00.0 On | Off |  
| 0% 37C P8 41W / 450W| 301MiB / 24564MiB | 0% Default |  
| | | N/A |  
±----------------------------------------±---------------------±---------------------+

±--------------------------------------------------------------------------------------+  
| Processes: |  
| GPU GI CI PID Type Process name GPU Memory |  
| ID ID Usage |  
|=======================================================================================|  
| 0 N/A N/A 1496 G /usr/lib/xorg/Xorg 59MiB |  
| 0 N/A N/A 1930 G /usr/lib/xorg/Xorg 70MiB |  
| 0 N/A N/A 2057 G …73947497,1085657924065483633,131072 153MiB |  
±--------------------------------------------------------------------------------------+

---

<div class="post-metadata">

**Author:** ![michael370](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@michael370](https://forums.developer.nvidia.com/u/michael370)\
**Post date:** [May 7, 2023, 2:06pm UTC](https://forums.developer.nvidia.com/t/cuda-12-1-tensorflow-on-linux-20-04-does-not-see-gpu-nvidia-4090-rtx/252514/2 "2023-05-07T14:06:17Z")

</div>

For the record…  
Torch sees it fine… print(torch.backends.cudnn.enabled) results in True

---

<div class="post-metadata">

**Author:** ![brami.martin0206](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@brami.martin0206](https://forums.developer.nvidia.com/u/brami.martin0206)\
**Post date:** [June 7, 2023, 7:54pm UTC](https://forums.developer.nvidia.com/t/cuda-12-1-tensorflow-on-linux-20-04-does-not-see-gpu-nvidia-4090-rtx/252514/3 "2023-06-07T19:54:52Z")

</div>

I seem to face the same problem. Keep me posted if you find a solution.

---

<div class="post-metadata">

**Author:** ![ar33bmuhammad](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@ar33bmuhammad](https://forums.developer.nvidia.com/u/ar33bmuhammad)\
**Post date:** [June 17, 2023, 4:59pm UTC](https://forums.developer.nvidia.com/t/cuda-12-1-tensorflow-on-linux-20-04-does-not-see-gpu-nvidia-4090-rtx/252514/4 "2023-06-17T16:59:18Z")

</div>

I’m pretty sure it has to do with the fact that the cuda 12.1 installation is only works with nvidia driver version 530.30.02 and not 530.41.03. I have the same issue.  
Please let me know how to solve it.