Even with out-of-core capabilities, pageing stuff in and out of disk (even the very fast SSD on MBP’s) is nowhere near unified memory bandwidth. Renderers are faster if they can keep high-res texture and geometry etc in memory. TM likes lots of memory, Ram and vRAM (although in the M1 case its all a single pool). On your hardware choice, I would say that 16Gb of unified memory is not sufficient IF you were to use something like Twinmotion. I’ve never used it so couldn’t comment on its integration interface or how up to date it is. Octane X is optimised for GPU rendering on M1 Pro/Max hardware and there is a Sketchup integration. Maxon Redshift is M1 native and utilises GPU rendering on M1 but no Sketchup integration. This is unlikely to change (ever) unless Apple implement ray tracing hardware to their GPU cores and a robust Metal raytracing API. Chaos only support GPU rendering via CUDA and RTX on Nvidia hardware (which is why I originally built a Hackinosh).

VRay is M1 native but does not utilise the GPU for rendering. Twinmotion is pretty much all GPU as its realtime and shares the same very fast memory pool as the CPU.Įpic / Twinmotion team say they are working on an M1 native version for release later in the year. Saving and opening large projects are significantly quicker on M1 due to the fast SSD. Twinmotion (although still running under Rosetta) seems to work very well for me, and apart from a small annoyance over the pointer click targets in relation to some UI elements (there seems to be a 10 pixel discrepancy), Output speeds are vastly quicker than my previous machine (i7 8700k, 64Gb RAM, Nvidia GTX1080Ti). I moved to a MacPro M1 Max (32Gb) from a desktop Hackintosh, I do a fair bit of visualisation and have used VRay and Twinmotion for a number of years.

Lumion macbook pro