I tested:
“C:\Program Files\MAXON\CINEMA 4D R19\Cinema4D.exe” -noopengl -render “Z:\octaneTest.c4d” -frame 1 1 -take “Main” -oresolution 800 600 -octaneGPU “0:true 1:true 2:false 3:false”
But it didnt work.
Deadline still set -nogui even with plugin configuration set to Cinema4D.exe, and didn’t work either.
Weirdly the next time I opened C4D it had set 2 out the 4 GPU active? I will try and test some more.
So after running
It renders using all GPUs but the once finished rendering it sets 2 GPUs to active. I also tried putting the -octaneGPU before -render but still the same result.
Hope this helps.
greetings colleagues!
did somehow solve this problem?
This should be resolved as of 10.1.16, are you having issues with your farm, or just checking to see how things are going? 
Yes, I have the same problem!
I contacted OTOY support and they confirmed the problem:
We are also having issues with our deadline test machines and currently investigating.
Please do let us know if you receive any guidance from the AWS team.
I would be very grateful for experience how to solve this problem and ready to conduct tests on my equipment according to your instructions
1 Like
I just have same issue, it seems like octane use all GPUs regardles worker GPU affinity settings (to be correct you need to use octane 2021.1.7 and higher - older versions has 2GPUs limitation).
Due my research, one worker with multiple GPUs has some benefits:
- you need one license per worker (octane/ C4D/ you name it)
- its easier to manage/ update
- you can build cheaper and more poverful setup with multi GPUs setup vs one latest GPU
I also run some test, but somehow its seems like 2 workers with 2GPUs are faster then 1 worker with 4GPUs (RTX 3070 to be exact)
1 Like
otoy technical support writes that this problem has not been solved
'in short, they have answered from a C4D point of view, but c4doctane is different, and you cannot control GPU assignment via commandline.exe, or cinema4d.exe, indifferently.
To add specific command-line Octane commands/arguments, the only way is to use Octane Standalone workflow."
1 Like