Blender 2.8 performance issues compared to C4D

Hello, dear fellow Blenderers.

As I’m transitioning more towards using Blender 2.8 instead of Cinema 4d for my 3D work I have started to notice some performance issues with Blender 2.8. Doing comparably similar work in Cinema 4D and Blender, I can’t help but notice that Blender starts behaving sluggish and that the viewport gets slow, even in seemingly simple projects.

Can I improve the performance?

My specs:
CPU: Ryzen 7 2700X
RAM: 16GB 3000MHz DDR4
GeForce GTX 1060 6GB

Could I update or add anything to speed up blender? Would I gain any performance increase from placing blender.exe on an SSD?

What I’ve noticed, is when using the subdivision surface modifier it can really slow down blenders viewport performance. Try enable the simplify check box in render settings it will speed up the viewport by lower the resolution of the subdivision surfaces. A SSD disk would only speed up loading of files. So if you use cache files (Alembic ,mdd etc) then it could help with a SSD, but I don’t think it will otherwise.

The subdivision implementation of 2.8 is a complete replacement of the one used in 2.79. performance isnt great yet but should be improved for the actual release of 2.80

2 Likes

do you subdivide intended for smoothing meshes or a garbage ones?

Unfortunately I’m not sure we’ll get further improvements to subdiv before 2.80 final. Recently a geometry cache was implemented which gets it back to something like the performance of the old un-accelerated blender subdiv code (this was 6x faster than it had been) but it’s still nothing like the accelerated OpenSubdiv implementation in 2.79.

A full GPU-accelerated version is planned, but this will not be until 2.81 (or later).

It would be most useful to know exactly what operations are slow. There are really no magic performance tunig options in most cases, and your hardware should be fine. Moving Blender to an SSD would help with startup type things, but probably not most activity.

A known issue is selection performance in large scenes, specifically ones with like 10,000 particle instances, where it can take several seconds to select an object in the viewport. This is under active investigation today by the dev team.

Subdivision Surface modifiers as noted are slower than they should be still. You can help this by turning down the view subdiv level to 1 (or 0) for interactive work.

Otherwise if you can describe what’s slow and how complex your scene is and if it seems to be correlated with anything, that would help in figuring out what’s going on.

1 Like