This might obviously be a C4D problem and not an Arnold problem but I have no idea.
With the IPR running, C4D will crash if you click to switch to another entire set of render settings. C4D will freeze for about 5 seconds after you click. then it will crash.
i.e. Running one render setting group for fast IPR preview with 720p, camera (AA): 2, diffuse: 1, etc.
then click to switch to another group for final render settings of 1080p, cam AA 7, diffuse: 9, whatever, etc.
If you can't get it to crash then try to keep clicking between render settings groups somewhat quickly.
I know this is a broad jab in the dark but I wasn't sure if it was a known problem or not.
Solved! Go to Solution.
Solved by peter.horvath6V6K3. Go to Solution.
I had not similar reports so far and I could not reproduce it, even when switching quickly. I tried with a basic scene though, C4DtoA 3.0.2 on Windows. Does it happen only in a specific scene or any scene? Maybe it depends on the actual settings?
I'm running C4D Education R20.059
C4DtoA 3.0.2.1 [4fb1c1c7]
Arnold core 6.0.2.1
Seems to be a specific scene that I have right now. It only seems to crash when I switch from my PNG 1080p render settings group to the PNG 720p group. Not the other way around, or to any other group settings.
I just did some more testing and found that it also rarely just freezes C4D completely but doesn't crash - and leaves my CPU maxed out at 100% usage. The only thing I can do is that it still lets me close C4D without having to force close it via task manager, but I can't reopen C4D because of some background(?) process that I have to force close.
I'll attach my scene for you.
Google Drive folder for scene files
I hope that the custom render settings actually transfer with the c4d files.
I have Octane plugin installed as well btw if that helps, even though I'm obviously not using it when C4D crashes with Arnold.
Thanks for the scene, I could reproduce the issue with it. Switching between render settings with and without a custom color manager causing the crash. It was actually reported before through a different channel and I already have a fix ready to be released in the next version.