Particle_Performance_Test

Simple Unity particle simulation, testing the performance of individual game objects vs drawing each particle to a single object.

To test performance, there are three Test States:

• Simulated - Simulates particles as custom classes- No gameObjects, only one Unity update loop running on the main handler. Requires gizmos to be enabled.
• As Game Objects - Simulates each particle as their own gameObject. This runs identical code, with each particle operating within their own update loop. Does not require gizmos.
• Rendered Objects - Simulates particles using the same structure as "Simulated," but attaches a simple SpriteRenderer to each one. Does not require gizmos.

Results

Viewing the framerate through Unity's stats at runtime, I was getting the following results:

• Simulated
  • Gizmos on: 280-300 fps
  • Gizmos off: 470-500 fps
• As Game Objects
  • Gizmos on: 240-260 fps
  • Gizmos off: 350-360 fps
• Rendered Objects
  • Gizmos on: 240-260 fps
  • Gizmos off: 370-380 fps

Simulated with gizmos off results in no visible particles, but comparing between Game Objects and Rendered Objects, there is a slight difference. At higher magnitudes of particles, the impact will likely be much greater, but it was interesting to discover that a single particle-handler, coordinating the updates and collisions for all particles, was more efficient than each particle running their own updates.

---

This was meant to be the beginning of a 2D liquid simulation for a game idea, but the project ended here. There were some fun discoveries, nonetheless.