Add to Render Queue/Send to Network

[mattjgerard]

If these both could internally use the Save Active Model Sets function, that would be fantastic and a huge time/space saver. So many product projects with 30+ models, and each one needs to be rendered out through a studio by itself. Right now it appears that sending to the render queue and sending to network saves the whole project for each studio. The archiving transfer and unpacking process would go so much quicker and require a lot less storage.

[DMerz III]

Agree with you here, I'm not sure if it is packing the whole file, but I think there are ways to optimize the "sending" part of the job.

If it is indeed packing the whole scene and all maps and geometry each time, perhaps the best option would be to put the studios interface as an option within the network rendering monitor?
That way you could send one package but select the studios in the queue somehow. This of course would mean any package is completely uneditable once it is "render ready".

You could lock all scene elements to changes and just be pure data? I don't know, I'm spitballin' here.

Ultimately, we lose a bit of efficiency when the job needs to "send" to the network and be status ready. The times we've hung our NR because we messed with our local keyshot application too soon is a kill on time. Removing the 'sending' process to a background process and not tied to the native Keyshot application is the only idea I have.

[mattjgerard]

I can almost certainly verify that it is saving the whole bip file, inactive model sets included. I had one project that was well north of 3gb with dozens of products in it, spent a week setting and sharing cameras, materials, environments etc, then when I went to submit about 50 studios to the NR, it took about 1-5 minutes per image to save and then another couple minutes to send, then another couple to unpack it on the NR master. Probably a good 5-10 miun for each image to actually start rendering once it was submitted. It took my computer out of commission for about 3 hours until the HD on the NR node filled up and it failed out of half the renders. Granted having a bigger HD in the render node is the easy answer for that, but it seems there should be some great effecineices to be gained somewhere.

I like the idea of setting the studios for render in the NR monitor. Then you are just sending the whole package once, not treating each one as a separate image. Even background rendering has this problem. Send something to the background renderer directly or through the queue and it packas the whole shebang together and sends it to itself.

I like brainstorming ideas, thanks David. Something to think about for sure!