Web Root: std::bad_alloc

### Custom Node Testing

- [x] I have tried disabling custom nodes and the issue persists (see [how to disable custom nodes](https://docs.comfy.org/troubleshooting/custom-node-issues#step-1%3A-test-with-all-custom-nodes-disabled) if you need help)

### Expected Behavior

Hi running the app after the last upgrade
running the app i have the crash with this errors:

i have 128Gb of ram
Nvidia 4090
and im on ubuntu...

what i can i try?

i already disabled custom nodes

### Actual Behavior

unable to run

### Steps to Reproduce

upgrade to latest

### Debug Logs

```powershell
`heckpoint files will always be loaded safely.
Total VRAM 24092 MB, total RAM 128081 MB
pytorch version: 2.9.0+cu128
WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for:
PyTorch 2.8.0+cu128 with CUDA 1208 (you have 2.9.0+cu128)
Python 3.9.23 (you have 3.10.14)
Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers)
Memory-efficient attention, SwiGLU, sparse and more won't be available.
Set XFORMERS_MORE_DETAILS=1 for more details
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 4090 : cudaMallocAsync
working around nvidia conv3d memory bug.
Using pytorch attention
Python version: 3.10.14 (main, Aug 8 2025, 22:51:23) [GCC 13.3.0]
ComfyUI version: 0.3.68
ComfyUI frontend version: 1.28.8
[Prompt Server] web root: /home/samuele/AI/ComfyUI/.comfyui/lib/python3.10/site-packages/comfyui_frontend_package/static
terminate called after throwing an instance of 'std::bad_alloc'
what(): std::bad_alloc
./run.sh: riga 3: 6434 Annullato (core dump creato) python main.py --listen
`

torch.version.cuda: 12.8
torch.__version__: 2.9.0+cu128
cuda.is_available: True
```

### Other

_No response_