RAM, CPU and Program limits

Does Blender have a limit on the number of objects, paths, keyframes, tri’s and vertices etc.? I ask because in my present Blender project, I wish to depict 10’s of thousands of star like objects (as if a particle system but more group or singularly tracked) Using my recent pc setup, I could produce about 9000 objects before impracticable slowdown. I went on to reduce my object template to 6 verts and 5 tri’s [a low res umbrella shape - to get some twinkling specularity]and cut everything back or off to stretch the object number to 12,500. I have 120 bezier circles operating as scaling and simutaneous rotation paths/platforms upon which I affix batches of objects which combined, produce a specific spiral track or obital system (hyperbolic). Blender will still produce 10’s (or even 100’s) of thousands of multifaced objects - even spheres in particle mode and I cant imagine the tri-count! yet 12.5k of “home-made” particles seems to be top whack - despite upgrading graphics card, and concomitant power supply. Doubling the Ram from 8 to 16 hasn’t helped either, but that may be an issue with the cpu. I don’t want to upgrade the cpu if I can avoid it but would like to know if Blender itself may have such limits. I don’t know that much about these things. It takes several weeks of redesign and plotting to expand the object number so a few solid days to find the max each time after redesign. I believe I am half-way to the desired 25k number but it is unclear what exactly may get me the other half. Any input graciously accepted.

To note, you can always divide the scene across multiple layers and only view one layer at once (enabling them all for rendering). This is a necessity for large scenes because Blender will otherwise slow down a bit.

Part of the issue here is that Blender has historically had a slow viewport compared to other apps. because much of it does not use modern OpenGL functions (many parts being just GL Begin > GL End), this is set to change sometime in the future however as the foundation for a modernized viewport is being done as part of the Viewport FX GSoC project. There’s also another project being done under that program that will multi-thread the dependency graph (in which tests have shown major speedups in scenes with a lot of motion paths and animation).

Even if these things solve your performance problems, you will still have the limit which is the RAM in your system (but for rendering at least can be mitigated to a certain point through things like instancing).

Moved from “General Forums > Blender and CG Discussions” to “Support > Technical Support”