iMac Pro

[DeckB]

I'm considering VRay for my office and, despite quite a few Bug Splats, have been pleased with it. Sounds like the Mac environment is not necessarily the best choice, but I'd like to get some advice to be sure the settings I'm using in VRay are 'best'.

First basic performance results. When rendering, the CPU option seems almost twice as fast as the GPU rendering. The CPU render also seem to be better quality as far as shadows/lighting. When interactively rendering, the speeds are closer. I assume the CPU's performance basically is the result of the number of cores in the CPU.

For GPU rendering, the 'C++/CPU' option in the asset editor must be selected. Does this mean the CPU is doing the majority/all of the work?

The other thing I've noticed is when selecting devices from the 'Tools' menu, the 'C++/CPU' option is not selected by default. There's also the 'error while parsing environment variable' (see attached), but I'm still able to select the CPU (assuming it's related to the error).

What would this selection affect (CPU, GPU, both) and should I be concerned about the error? There doesn't seem to be a difference in performance when it's selected.

Thanks!

[DeckB]

Does the 'Select devices for V-Ray GPU rendering' only apply two render server, and is that different that the standard render process?

[robD]

V-Ray does not currently support gpu rendering via opencl on amd cards, only on nvidia cards using cuda. V-Ray will execute cuda code on the cpu (c++/cpu) option but obviously at the same render speed (or slower) as straightforward cpu render. In you devices window it pretty much says same. 2 devices, 1 (viable) platform. The amd card can?t be used for gpu render until vray supports opencl (which actually may be an Apple issue but I might be wrong

[DeckB]

Thanks, robD.

What's interesting is that the default selection in that window is the GPU, so that window must just identify GPUs even if it cannot use them. I can select the CPU (like I did in the first post), but it deselects on close.

So technically all of the Vray rendering on an Apple with AMDs is CPU?

Like I mentioned, a final render is significantly faster via CPU than using GPU & C++, which supports what you're saying.

[robD]

yep, all CPU with AMD cards, you could add an eGPU enclosure with an Nvidia card in it which would provide you with GPU / Hybrid rendering using CUDA and the AMD card just drives the display until there is an OpenCL implementation.
Personally I've built a Hackintosh with Nvidia 1080Ti for rendering.

[DeckB]

Awesome, thanks again robD!