Skip to main content

Installation

Prerequisites​

  • Miniconda or Anaconda
  • Git
  • NVIDIA GPU with CUDA support (for GPU acceleration)

Step 1: Clone the Repository​

git clone https://github.com/Markeloen/AtmoTurbSim.git
cd AtmoTurbSim

Step 2: Create the Conda Environment​

conda create -n atm python=3.9 -y
conda activate atm

Step 3: Install TensorFlow with CUDA Support​

TensorFlow versions below 2.11 are required for native GPU support on Windows:

pip install "tensorflow<2.11"

Step 4: Install Dependencies​

pip install -r requirements.txt

Step 5: Verify GPU Access​

python -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"

You should see output like:

[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]

If the list is empty, check your CUDA/cuDNN installation.

Troubleshooting​

No GPU detected​

  • Ensure NVIDIA drivers are up to date
  • Verify CUDA toolkit is installed and matches the TensorFlow version
  • Check cuDNN is installed in the correct CUDA directory
  • Run nvidia-smi to confirm the GPU is visible to the system

TensorFlow version conflicts​

The simulator requires TensorFlow < 2.11. Later versions dropped native Windows GPU support in favor of WSL2. If you need a newer TensorFlow, use WSL2 on Windows.

Import errors​

Make sure the conda environment is activated before running:

conda activate atm