FWIW on a 6000 Pro it takes 68 seconds for 10 seconds 480p video, (cold) same demo workflow as you used. Set "megapixels" to 2.0 (1920x1088) and same video seems to take 5+ minutes, not sure if everything is right/correct at the moment.

As far as I can tell, the current ComfyUI nodes don't even do compilation, and I haven't looked into what attention mechanism they're using, but I'm sure with time these durations will come down even more.

I must be doing something incredibly wrong.

I'm on an RTX 5090. I told it to make a 5 second 864x480 video, it's been running for over 30 minutes and is only 35% done in the SamplerCustomAdvanced step.

EDIT: Oh, I'm an idiot. Forgot I had a llama.cpp webserver running with a model loaded. Killed it and it finished very fast.