V-Ray RT with NVidia VCA

I am terribly skeptical of these VCA boxes. I use GPU rendering extensively and find it to be most useful when you can get IMMEDIATE results as you build up your scenes – that does not include sending data 30seconds to network to render a simple watch and then 30 seconds again once you add/modify your geometry. This is a watch, now imagine 10-100x more complex scene. It kills all interactivity. Simple mid range gaming card will give you better visual estimation within that time it took to send this data. Animation – yes, absolutely! However again you cannot use it so simply in production because these boxes are rigid in their setup running linux. You will have plugin errors, version mismatch – it just wont work. Having several Nvidia GPUs locally you have all the control in the world and flexibility to use them for rendering, simulation or anything else. This here seems more of a “limited renderfarm” solution for very specific audience. They refer to it as “visual effects solution” but the GPU rendering lacks much of the features and flexibility needed for these “visual effects”. I hope to see vrayRT mature a bit faster to support more of the CPU renderer features as well as Octane to finally integrate passes – now that would be amazing and affect us all.