Question Optimize moving particles in Update

In my game i have multiple planes where i move single particles(stars) to achieve a Starfield with parallax effect.
Since its about 1k particles, it contributes about 80% of my update time of all my monobehaviours.
What are ways to optimize this code?
Would Unity Dots be a good case for a problem like this?

Code (CSharp):

1. private void FixedUpdate()  
2. {    
3.     for (int i = 0; i< MaxStars; i++)    
4.     {
5.         Vector3 pos = Stars[i].position + transform.position;
6.         pos.y += -CalcedYShift;
7.         if(pos.y < transform.position.y - yOffset)      
8.         {
9.             pos.y += FieldHeight; pos.x = Random.Range(0, FieldWidth) - xOffset;      
10.         }
11.         Stars[i].position = pos - transform.position;
12.     }
13.     Particles.SetParticles(Stars, Stars.Length);
14. }

 

Can it just be a particle system? That's how most of the kool kids do their starfields.

 

There's also the Bethesda way of doing starfields. Not recommended!

I hope you're not using a ParticleSystem each with one particle?

Like Kurt mentioned, a single particle system that emits "star" particles on one side (vertical or horizontal) and moves them linearly with minimally varying speeds to the opposite side would be sufficient. It works with practically zero code even and it'll be fast even with 1000 particles.

 

Actually I did make a custom starfield for my Jetpack Kurt, as I needed a specific easily-movable volume affected by varying winds, aware of FixedUpdate physics. It is purely to visualize "wind flow" right in front of the player, regardless of which direction they might look or turn their head during play.

I didn't see a handy way to do that in the ParticleSystem, so this was my result:

https://gist.github.com/kurtdekker/e3117bef40fbf329d87c2bfd3c0121d2

The key takeaway of that script was to do as little work as possible in FixedUpdate().

Here it is in action:

 

What's the complete code of the script? Is it just

FixedUpdate

with some fields?

 

there was indeed no reason to do the star moving manually, the behavior update reduced from 6 to 1.5ms.

The particle system itself is barely noticable with 0.15ms

thanks!