I know the gaming world has tools to optimize performance for their target devices... Do we have something like that in the Blender realm?
I would love to understand my blend files better with an eye toward wringing as much out of them as possible at render without breaking the bank or melting a CPU -- how to stay under GPU memory, where and how mesh densities are killing times, how and where trade-offs exist between skin sizes and render speeds.... Is anything like this available?
I suspect this question also entails stuff I do not know to ask or look for -- when skin sizes could be higher/lower in a scene, using LoD more effectively than I do... and more, as I have only an experience-based sense for when adding or subtracting "weight" helps... and that's not really scientific.
We (try to) do high-realism scenes with natural environmental components -- lighting, volumetrics, particles, water... and optimizing for hardware rendering is proving to be an important issue, especially as me move to 4K stuff.
I know (some) tricks to optimize scene weights: We use scrims, layer passes, post processing of volumetrics and more... but if I could analyze a live file to see where render demands are spiking, that would be cool. So this is less a question about render optimizing approaches in general and more about understanding the guts of render demands on a per-file basis.