Comfyui not using GPU, how to fix this?
We are running in instance of Comfyui and it's using CPU and not GPU, using a 4090. How can this be fixed?
MCheck if your pod is not running in 0 GPU Mode
LIt is running on a 4090.
JYep
JWhat's ur comfyui args
Mbtw I just made better comfy ui template that not breaks on extensions instalation
#Better ComfyUI - CUDA12
#Better ComfyUI - CUDA12
DI don't think that GPU utilization metric is 100% accurate and it also not real-time, better to use
nvidia-smiWhen using the Better Comfy Setup and I connect to port 7777 it says "GPU 0 NVIDIA GeForce RTX 4090"
JYes
Jthats normal... whats your point
LI don't know. It's still using the CPU. I'm not sure where to look for the comfy args, other then using the log on Runpod, but there is nothing there that would indicate anything useful.
Jwhats your workflow like
Jcheck the comfy docs
JWhy do you think it's using the cpu??
LWorkflow works fine locally
LBecause I have the crystools Nodepack and it shows it's using the CPU
LAnd it takes forever. On my local 4090 it's done in 10 seconds.,
L
Jcreate a new pod maybe that pod has a problem?
LAlready tried that
Jreport it to the contact support if it does
DDo you see anything in the ComfyUI logs?
Llogs.txt43.52KB
DThis looks fine
DNot sure whats with xformers not being available tho, but that shouldn't cause it not to use GPU.
DIt could impact generation times tho.
DDoes your local have xformers installed?
D"pytorch version: 2.3.1+cu121" this could be a problem for xformers, there is no compatible xformers for torch 2.3.1.
Jtry your local files to a pod if you want to replicate it
Jand install the same packages as your local
Dcan't install xformers if its installed on local
Dhave to reinstall torch
Donly 2.3.0 is compatible with xformers not 2.3.1
LI don't think my local instance has xformers installed but I'm not sure. it's not in the start up file.
Dactivate local venv and then:
Dthen you can compare the versions
Dotherwise also
Jwow didn't know pip freeze exists its cool hahahah
LOnce I can find terminal in this new backend you got it
JHuh?
Jwhich backend did you mean
LIt's not a jupyter notebook so finding the terminal is more difficult. THe port 7777
Dcode-server
Johh
Jtry to screenshot the left menu bro
Jmaybe i can help pointing it out
L
JNo other bottom ones?
Ji thought there are like 3-4 buttons on the bottom left?
L
