OTOY Forums

I purchased 1 C4D license and 2x Standalone so that I could render on my 2 machine (1 has 2 x 1080s and my main machine has 3x 1080ti) both machines are watercooled, one with hyrbids and one with full watercooled.

When testing render with network rendering, I get pretty much the same render times, sometimes using network rendering actually slows the renders down...

I'm not sure why this is, but surely if Octane used the other machine to render different frames...Somehow doing a calculation of both machines hardware then sharing out and distributing the frames accordingly so that the additional (less powerful) machine can actually use it's full potential.

At the moment I'm seeing no benefits on having 2 machines, I did think I could load the scenes on my slave machine but then I would have to purchase another Octane C4D license...Which is not an option for me at the moment.

Hi mikeadamwood,
which exact version of plugin are you using?
What kind of network connection/speed do you have?
Please try with a scene that has a couple of minutes of rendering time on the Master only.
Note that you can export from c4doctane in .orbx animated mode, and render it with Standalone on another machine.
ciao Beppe

bepeg4d wrote:Hi mikeadamwood,
which exact version of plugin are you using?
What kind of network connection/speed do you have?
Please try with a scene that has a couple of minutes of rendering time on the Master only.
Note that you can export from c4doctane in .orbx animated mode, and render it with Standalone on another machine.
ciao Beppe

Sorry for late reply!

I'm on latest version of Octane 4(Not 2018) a gigabit Ethernet and when I do the test to compare I get pretty much same result...If not sometimes longer when adding slave...Would I notice a different in longer renders you think?

I've attached a screenshot of a test, highlighted in green is the network render with all GPUs binded....Highlighted in red is the master only...



Just seems odd that I've getting literally no speed increase at all after investing in another machine :/

Hi,
please, try to decrease the Max Tile sampling at 8, or lower, in the Kernel settings, and report back, thanks.
ciao Beppe

Also, frames that only take a few seconds to render on one machine may not see much improvement with network rendering. Depending on the scene, plan on at least a few seconds of network overhead and coordination for every frame. Try rendering frames that take a minute or two each, and you should see a big difference.

If you have lots of individually fast (less than 15 seconds each, for example) frames to render, it may be significantly more efficient to split the job into two parallel tasks on two machines instead of trying to shoehorn them all through a single, networked render queue.