R
RunPod•5mo ago
smoke

Estimated time comparison - Comfy UI

Hi everyone, I've been looking at the various different GPU options for serverless and I am trying to see if anyone has a rough estimate of how many times each GPU would be faster/slower and if this is even possible to calculate. There is no actual formula obviously but I am wondering if someone has similar experiences. In my case, it takes around 355 seconds to run my workflow on my local machine (GTX 3080 Ti). Is there a rough estimate on how long this would take on the serverless GPU's? I assume that it will be faster. I will compare all the GPU's soon anyways but I was just wondering.
4 Replies
ashleyk
ashleyk•5mo ago
Too many variables, mostly delay time and cold start time.
smoke
smoke•5mo ago
I see. Do you think any of the serverless GPU's will be faster than my local machine though? I assume so but I just want to make sure. I'll try it soon anyways.
ashleyk
ashleyk•5mo ago
Your local machine probably doesn't incur cold start times so it depends....
smoke
smoke•5mo ago
That's true. Okay thanks a lot! I'll just try it out.
Want results from more Discord servers?
Add your server
More Posts
Zero GPU issueI wanted to start up a SD Comfy UI pod I created the other day and when I start it up I get a pop-upStart and stop multiple podsI have a product that will allow users to submit video editing requests that can range anywhere fromAny plans to add other inference engine?Hi I'm using vllm worker now but when it comes to quantized models vllm works poorly. Too many vram Are there any options to retrieve container logs via API?Need it for monitoring purposes.Serverless scalingI'm considering using runpod for commercial use. I need reliable, relatively cheap scaling for this "Failed to return job results. | 400, message='Bad Request', url=URL('https://api.runpod.ai/v2/gg3lo{5 items "dt":"2024-02-19 02:45:23.347011" "endpointid":"gg3lo31p6vvlb0" "level":"error" "message":"`runpodctl send` crawling at <1MB speedsHi there! I'm a big fan of RunPod for training SDXL, and have spent a bunch of time (and money!) iteStable Diffusion API Execution TimeI am posting this for a response from runpod support @flash-singh or anyone other than @justin Is Cannot create pods even there are available gpusWhen I query as following, ```graphql query { gpuTypes(input: {id: "NVIDIA L4"}) { id Transfer/Duplicate Network VolumeAs per title, can I duplicate my network storage or temporarily move it to a different region so tha