Comfy-Org / Comfy-Org/ComfyUI

Confi UI KSampler issue with CUDA

Open
#6,561 10 comments 0 reactions 0 assignees View on GitHub
Potential Bug
Dominant language
Python
Stars
133k
Forks
15.7k
Avg merge
1d 7h
Merged PRs (30d)
158

Description

### Expected Behavior

Hello,

new to the confi-UI world,

got this error message after updating my nivdia driver for the RTX A5000 24gb ( 4 gpu
system is : win11 ( 10.0.22631 build 22631) pro AMD Ryzen Threadripper PRO 5975WX 32cores 360 Mhz, 64Log ... etc etc system.
GPU driver is version 32.0.15.6603 ( from win sys info ) from Nvidia Control panel > 566.03 , driver type DCH, direct 3d 12_1 Cuda cores : 8192.

here is the error message :

No operator found for `memory_efficient_attention_forward` with inputs:
query : shape=(1, 4494, 24, 128) (torch.bfloat16)
key : shape=(1, 4494, 24, 128) (torch.bfloat16)
value : shape=(1, 4494, 24, 128) (torch.bfloat16)
attn_bias :
p : 0.0
`fa3F@0.0.0` is not supported because:
xFormers wasn't build with CUDA support
requires device with capability > (9, 0) but your GPU has capability (8, 6) (too old)
operator wasn't built - see `python -m xformers.info` for more info
`fa2F@0.0.0` is not supported because:
xFormers wasn't build with CUDA support
operator wasn't built - see `python -m xformers.info` for more info
`cutlassF-pt` is not supported because:
xFormers wasn't build with CUDA support

![Image](https://github.com/user-attachments/assets/1e19c555-5356-4325-9e65-bd161566dc42)

as a file on the side, i put this PNG workflow ( ComfyUI_00593_.png ) that was working the day before this " SERIES OF UNFORTUNATE EVENTS :))) " .

and if by any means if this is usefull > i also put the SETUP that trigger this chain of events for updating Nvidia drivers and CUDA, and installing extra nodes inside Confi-UI that might have create CONFLICT ( as some of them were pointing the possible issue .... ) > " ControllableCharacters_v03.json " and " Pose_sheet_v02.png "

[ControllableCharacters_v03.json](https://github.com/user-attachments/files/18503609/ControllableCharacters_v03.json)

![Image](https://github.com/user-attachments/assets/18434d78-7f28-4bfe-991f-0988bffae258)

### Actual Behavior

the error message again :

No operator found for `memory_efficient_attention_forward` with inputs:
query : shape=(1, 4494, 24, 128) (torch.bfloat16)
key : shape=(1, 4494, 24, 128) (torch.bfloat16)
value : shape=(1, 4494, 24, 128) (torch.bfloat16)
attn_bias :
p : 0.0
`fa3F@0.0.0` is not supported because:
xFormers wasn't build with CUDA support
requires device with capability > (9, 0) but your GPU has capability (8, 6) (too old)
operator wasn't built - see `python -m xformers.info` for more info
`fa2F@0.0.0` is not supported because:
xFormers wasn't build with CUDA support
operator wasn't built - see `python -m xformers.info` for more info
`cutlassF-pt` is not supported because:
xFormers wasn't build with CUDA support

### Steps to Reproduce

[ControllableCharacters_v03.json](https://github.com/user-attachments/files/18503638/ControllableCharacters_v03.json)

![Image](https://github.com/user-attachments/assets/eb55f651-c5d2-46cb-be2f-e0f33b405c04)

### Debug Logs

```powershell
No operator found for `memory_efficient_attention_forward` with inputs:
query : shape=(1, 4494, 24, 128) (torch.bfloat16)
key : shape=(1, 4494, 24, 128) (torch.bfloat16)
value : shape=(1, 4494, 24, 128) (torch.bfloat16)
attn_bias :
p : 0.0
`fa3F@0.0.0` is not supported because:
xFormers wasn't build with CUDA support
requires device with capability > (9, 0) but your GPU has capability (8, 6) (too old)
operator wasn't built - see `python -m xformers.info` for more info
`fa2F@0.0.0` is not supported because:
xFormers wasn't build with CUDA support
operator wasn't built - see `python -m xformers.info` for more info
`cutlassF-pt` is not supported because:
xFormers wasn't build with CUDA support
```

### Other

_No response_

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.