I started it around noon, then did some running around until about 3, Blender crashed after I left. (My system uptime was over 60 days at that point, so that's my fault)

I started it again from a couple frames before it crashed, an hour later I went on a date and got back around 9. As I was about to leave I saw a text from my roommate saying that we lost power at 4:30.

When the power came back at 10ish I started it again.

Only about 2000 frames to go out of the 14472 frames in the render. And at about 1.25 seconds per frame, It shouldn't be too much longer.

If I didn't have to have this done before tomorrow (I've got a couple projects that need to finish this week) I would have just had round 3 go while I was sleeping and dealt with any potential problems (like another crash) in the morning.

tis why at minimum a proper computer lab has 5 machines: router, firewall, LAN, servers, & work.
Renders should be left to either LAN or servers to work, while you game.
gamers hate being told you don't proper computer labbattlestation without at the least 5 machines running. Routers, firewall, and, servers should be up 24/7.

I'm pretty sure there has been a kernel update in the last 60 days, which could make your renders 0.2s/frame faster. Also, in my case using blender cli also renders a little faster.

Unfortunately before my NAS got it's much needed upgrade to be for offloading rendering tasks to, component prices went to the moon.

I run Kubuntu so my kernal isn't exactly "at the forefront" of development lol

I didn't know about Blender CLI before your comment, I might have to take a look at it in the future for some other projects that'll inevitably take awhile.

This one took so long because it was 1 shot of 14472 frames, my next version of it is def gonna take longer to render.

That's bs&ʧ problem. The’re datacenters waiting to be properly distributed to the needy. You only need a party of 5 to cut off a DC, take a rack, and serve yourselves.

If your using your machine for work I'd reccomend getting a UPS so it can keep running during short power outages and save having to restart render jobs.

I've never done anything that long with Blender, so I have no experience, but I could also imagine blender CLI being a little more reliable in terms of crashing.

Or at least when it does crash, it'll be more obvious. I assume someone out there has made some kind of "automatically restart it from where it left off if it crashes" script, if the CLI doesn't do that itself 😛

I don't use blender, but I got curious anyway after the other person's comment. Obviously I haven't used it, but you might look into the repo, might make it easy for you to auto-restart crashed jobs

https://github.com/dmtvision/Blender-Auto-Render

this is why we justify buying a handheld 😃

Just never render videos. You always render individual frames which you then turn into a video. You won't lose progress that way. But probably everyone rendering animations knows that.

Second GPU dude. Just sell your house and add a second GPU.

I was rendering the individual frames which is why it went so smooth in spite of all the issues

I learned that lesson long ago lol

The Steam Deck is more and more tempting by the day

My homies and I about to upgrade our homelabs

ggwp

EDIT : now it's time to put the frames into a video, and render that

That went smooth as silk with a render time of less than 0.1 seconds for every frame

I would just throw a small 10 second clip from it right here in the comments just for some reason this instance (blahaj) doesn't let me 😿

I'd have to upload it somewhere else

The instance I'm normally on (slrpnk) is down due to a sudden hardware failure a couple days ago and won't be back until December