"Torch not compiled with CUDA enabled" means the PyTorch installed in your ComfyUI environment has no CUDA support. The fix in the ComfyUI README: uninstall torch and install it again with the CUDA command.
Fix (NVIDIA)
Inside the environment ComfyUI runs in:
pip uninstall torch
pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu130
Then confirm PyTorch sees the GPU:
python -c "import torch; print(torch.cuda.is_available())"
torch.cuda.is_available() returns a bool indicating whether CUDA is currently available. If it prints True, start ComfyUI again.
Make sure it is the right environment
A common cause is installing into one Python environment and running ComfyUI from another, so activate the same environment for both. The README also says to use a cu130 or newer PyTorch on newer NVIDIA GPUs, and to update your NVIDIA drivers if the portable build does not start.
To confirm which device ComfyUI chose, see how to check if ComfyUI is using the GPU. A fresh install is in how to install ComfyUI on Linux, and the ComfyUI custom node install guide covers nodes that pull in their own dependencies.
Sources
- ComfyUI README: NVIDIA and Troubleshooting. github.com
- PyTorch docs: torch.cuda.is_available. docs.pytorch.org