Hello, I have a scene with 36 xmesh loaders in it. When I render it locally with Mental Ray it translates the scene super fast and then renders the frame, however, when I send it to the render farm the translate time for the xmesh loaders goes through the roof and the scene does not actually start rendering for an hour or two (maybe more, I never let it complete, it took 20 mins to translate 4 of the xmeshs in the scene).
I feel like I did something wrong here when I sent it to the farm but I couldn’t really see anything super obvious. Does anyone else have any suggestions or run into this before that can shed some light on this?
Thanks! I tried this on one machine and it didn’t seem to make a difference, the translate time was still taking a long time compared to the local machine.
I am going to try it with v-ray instead of mental ray and see if that makes a difference, also I have the mesh loader plugin being loaded from a central plugin folder off the server, could this be a factor as well?
The location of the plugin itself should not have any effect, but the location of the XMesh data could. XMesh uses multiple files to read compressed binary channels in, and reading files over a network can be slow depending on how the network was set up (we had this issue when we switched from Fusion to Nuke at Prime Focus back in the day as our network was optimized for small files, but Nuke wanted optimization for large files due to the way it was accessing image data).
So the first thing I would suggest is this:
*Copy the XMesh assets to a folder on the local drive of the network machine
*Submit a job with the same local path set in the XMesh Loader
*Compare the time
If it renders fast, then you know your network access to the assets is the issue.