Network render times...

[pixelrocket]

I recently added a standalone license to my current Octane workflow. I have no problem seeing the slave and getting it to show up and render inside Cinema 4D. In my current setup, I am using my Mac as the master machine and I have another PC as the slave. For the current round of tests, I am using 1 GTX 980ti in the Mac, and 1 GTX 780 6GB in the PC. Cinema 4D version 16.050, Octane 2.24.2, Octane C4D plugin V 2.24.2 R2. Both computers are connected through a Gigabit network.

If I use Octane standalone, the render times are 1:22 with just the Mac, and :49 when adding the network rendering slave. The second machine is obviously making a difference when rendering with the standalone version.

Standalone.

SA.png (6.58 KiB) Viewed 2939 times

Standalone & Slave.

SA Network.png (6.42 KiB) Viewed 2939 times

However, if I render through the Picture Viewer in Cinema 4D, or through the render queue, the difference in render times using the network render machine is much less. When rendering the file through Cinema 4D, with the Mac only, the render time is 1:33. If I add the network render machine, the render time is 1:12. While the extra network machine does help some, it's not anywhere close to the performance when running the scene file through Octane standalone.

Cinema 4D render queue.

Render Que.png (10.14 KiB) Viewed 2939 times

Cinema 4D render queue with Slave.

Render Que Network.png (10 KiB) Viewed 2939 times

For these tests in Cinema 4D, I used "resend only updated objects" and the current scene file I am testing with is pretty small with just a handful of objects. Depending on the scene file, using the network slave through Cinema 4D sometimes results in a much longer render time, sometimes double what I would get if I rendered locally.

I have saved the log files for each session on the slave but I don't really know what to look for to see what could be causing the time difference. I also have not run any network speed tests since the render times were so different between the plugin and the standalone version.

Has anyone else run into this speed difference? Is the speed difference just a result of using the standalone vs Cinema 4D plugin for rendering?

[aoktar]

how much performance gain do you see just in Live Viewer with/without network?

[pixelrocket]

Using the live viewer, the times are better, but still not as good as the results from the standalone render.

Live viewer without network render:

Live viewer with network render:

[aoktar]

Using the live viewer, the times are better, but still not as good as the results from the standalone render.

Live viewer without network render:

LV.png

Live viewer with network render:

LV NR.png

SA vs. plugin different is by render priority.
can you try with disabled priority and 8bit render buffer?

[pixelrocket]

I am not using render priority on the Master machine since it has a dedicated card for Octane already.

Ah, once I compared tone mapped images from Standalone and Cinema 4D, render times were similar.

Thanks.

[aoktar]

Can you please test it on PV with 8bit render buffer from settings? Thanks!

[pixelrocket]

Ok, I rendered the same file to the picture viewer with 8 bit render buffer.

First render time is without network rendering, second render time is with network rendering.

Screen Shot 2016-01-23 at 7.06.40 PM.png (11.95 KiB) Viewed 2831 times

[aoktar]

Here is my test results. I take the times from beginning of first sample to maxsample. It's very simple scene on PT, maxsamples=1000, fullHD