ComfyUI not starting, "CUDA error: operation not supported"

### Expected Behavior

ComfyUI does not start up at all. CLI outputs an error

* ComfyUI 0.3.14 portable
* Windows 1
* NVidia A16-8A, driver version 528.24, driver type DCH

### Actual Behavior

ComfyUI does not start up

### Steps to Reproduce

double click the `run_nvidia_gpu.bat` file

### Debug Logs

```powershell
c:\Users\username\AppData\Local\ComfyUI_windows_portable>.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build
Checkpoint files will always be loaded safely.
Traceback (most recent call last):
File "c:\Users\username\AppData\Local\ComfyUI_windows_portable\ComfyUI\main.py", line 136, in
import execution
File "c:\Users\username\AppData\Local\ComfyUI_windows_portable\ComfyUI\execution.py", line 13, in
import nodes
File "c:\Users\username\AppData\Local\ComfyUI_windows_portable\ComfyUI\nodes.py", line 22, in
import comfy.diffusers_load
File "c:\Users\username\AppData\Local\ComfyUI_windows_portable\ComfyUI\comfy\diffusers_load.py", line 3, in
import comfy.sd
File "c:\Users\username\AppData\Local\ComfyUI_windows_portable\ComfyUI\comfy\sd.py", line 6, in
from comfy import model_management
File "c:\Users\username\AppData\Local\ComfyUI_windows_portable\ComfyUI\comfy\model_management.py", line 166, in
total_vram = get_total_memory(get_torch_device()) / (1024 * 1024)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "c:\Users\username\AppData\Local\ComfyUI_windows_portable\ComfyUI\comfy\model_management.py", line 157, in get_total_memory
_, mem_total_cuda = torch.cuda.mem_get_info(dev)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "c:\Users\username\AppData\Local\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\cuda\memory.py", line 738, in mem_get_info
return torch.cuda.cudart().cudaMemGetInfo(device)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: CUDA error: operation not supported
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
```

### Other

_No response_