Scene renders fine on local vray standalone, does not render in chaos cloud

[elivnA]

Hello,

This problem is related to a scene that takes way too long to render on my local machine, and Distributed rendering doesn't work either (renderslave renders black for some reason). So I diverted to Chaos Cloud, only to have it not render at all. Could you help me figure out what the problem is? I submitted serveral tickets and didn't get much of an answer (or a fast answer), so I will try here. I have a downloadlink ready with an archived maxfile. Important to note is that this image (with about 15 fewer buildings) rendered fine the last 2 times in Chaos Cloud.

I suspected it was a memory issue, because there are about 70 Revit buildings placed using vrayproxy, each 1-10 mil polygons. There are about 15000 Evermotion trees spread with Forestpack, and about 6000 instanced vray spherical lights. After reading Lele's memory optimisation tutorial I did the following optimisations:

• set all bitmap filtering to 1
• replaced normal bitmaps with vraybitmaps
• set the GI to BF/BF
• render with bucket image sampler
• set the bucket size to 16
• turned off memory frame buffer
• removed some unneeded geometry, but due to time limitations and the scope of the project I had to stop. There is much to be gained here geometry wise.
• set the number of reflections/refractions to 2
• turned on "conserve memory"
• set dyn. memory limit to 58000mb (my workstation has 64)
• removed unneeded elements (currently using 3: zdepth, denoiser and cryptomatte)

Please see below for a lowres preview rendered on vray standalone, on my local PC:

• 

[nikoleta.garkova]

Hi elivnA

Regarding not receiving a reply from Chaos support/getting one after a while - we had an issue with our ticketing system and we didn't get some messages on time. Apologies for the inconvenience.

That being said, it appears that you've opened two tickets on our system - you should've already gotten a reply. We would need to get the scene files (both versions - original and the optimized one) so we could investigate them better. Thanks for the understanding.