Apologies if this has been queried before but I couldn't find any answers in the forum.
We're just starting to test out network rendering in our company and so far everything is working, I can see the slave PCs running 2.25 and they are picked up when rendering.
However, I've noticed that while adding more cards speed up the rendering, the amount of time that Octane spends saving passes stays the same. If anything it actually manages to save passes quicker when I'm only rendering locally because I'm assuming that it doesn't have to transfer any data over the network.
Is this correct? Or should the slaves actually be speeding up creation of passes as well?
Many thanks.
Network rendering and passes
Moderators: ChrisHekman, aoktar
In version 2, the render results are stored on the GPUs of the slaves. When you save them out, the master needs to fetch all the results and blend them into one image, before it can save the image. This needs to be done for every pass and is quite expensive to do.roarke80 wrote:Apologies if this has been queried before but I couldn't find any answers in the forum.
We're just starting to test out network rendering in our company and so far everything is working, I can see the slave PCs running 2.25 and they are picked up when rendering.
However, I've noticed that while adding more cards speed up the rendering, the amount of time that Octane spends saving passes stays the same. If anything it actually manages to save passes quicker when I'm only rendering locally because I'm assuming that it doesn't have to transfer any data over the network.
Is this correct? Or should the slaves actually be speeding up creation of passes as well?
Many thanks.
In version 3, the render result is accumulated on the master while rendering and when you save the passes, no data transfer from the slaves needs to happen.
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra