Particle grass vs. multimesh grass - which performs better?

Archive
1
:information_source: Attention Topic was automatically imported from the old Question2Answer platform.
:bust_in_silhouette: Asked By Macryc

Guys, those of you who’ve tried both particle shaders for grass as well as multimesh grass, which did you find more performant?
I thought that particle grass worked in a similar way to multimesh, ie, all parts are really one mesh with a single draw call etc, but I notices horrendous frame drops with particle grass, even with seemingly low numbers of particles (like 5000, which is nothing for grass). I do recall with multimesh I was getting much better results in 3.4 (now using 4.0).
Any insights will be appreciated.

What happens if you spawn 5000+ Particles with a regular particle processing material? If this also has frame drops I would think it’s more of a general particle system problem than a grass problem.

It could be that reading back data from the GPU is more intense in 4 than in 3 due to it being not yet optimized and thus it performs worse in 4. EDIT: Nevermind this probably works entirely different in 4 so that assumption is void.

omggomb | 2022-06-24 06:50

2
:bust_in_silhouette: Reply From: Inces

Multimesh should be better option than particles. It has the same base mechanism, but less options and gimmicks. Its shaders are far less complicated. You still need to optimize it though, like hide visual instances that are offscreen.

I’ve tested with the same number of meshes and same meshes and 3d particles seem to be performing a little better although it’ a tiny difference, like 2-4% better.

Macryc | 2022-06-29 10:23