I have successfully configured TensorFlow on a fresh Ubuntu 22.04 installation, following the instructions provided in the official TensorFlow documentation link. I have setup CUDA® Toolkit 11.8 cuDNN SDK 8.6.0 in my conda enviroment and I have my nvidia drivers installed. My system is using a Nvidia 1080 GPU, and I've verified that TensorFlow can detect and utilize the GPU in both the terminal and Visual Studio Code (VSCode).
However, when I attempt to use TensorFlow within DataSpell, the Jupyter server fails to detect my GPU. Here's the output I receive when running the following code snippet in DataSpell:
import tensorflow as tf
print(tf.config.list_physical_devices('GPU'))
Output in DataSpell:
I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find CUDA drivers on your machine; GPU will not be used.
I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find CUDA drivers on your machine; GPU will not be used.
I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations. To enable the following instructions: AVX2, FMA, and others, rebuild TensorFlow with the appropriate compiler flags.
W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT.
[]
W tensorflow/core/common_runtime/gpu/gpu_device.cc:1960] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and set up the required libraries for your platform.
Skipping registering GPU devices...
However, in both VSCode and the terminal, I get the following output when running the same code:
Output in VSCode and Terminal:
I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2, FMA, and others, rebuild TensorFlow with the appropriate compiler flags.
W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT.
[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]
I expected the code in DataSpell to return an array of available GPUs, similar to how VSCode and the terminal behave. This is a fresh install of Ubuntu and I have configured a conda enviroment for tensoflow following the official documentation. Only dataspell does not detect the gpu.
