Performance Question: MoverConstraint or CFrame

[WillyEdison]

Hello there!

I’m optimizing our team’s game for lower-end devices, and I have a question about the best path to take. Our lobby ship has a giant propeller in the back that spins. Currently, it adjusts CFrames every frame, but with the new AngularVelocity mover constraint, would that be more efficient? Or is modifying the CFrame every frame better?

I’m personally curious, if one of these options is better, what makes it better? Does one leverage less CPU power? Is there an advantage (or disadvantage) to using physics-based solutions like MoverConstraints? The more details, the better! I’m curious about the implementation.

Thank you for any help!

[dthecoolest]

@WillyEdison

Our lobby ship has a giant propeller in the back that spins

For this situation I believe it shouldnt matter either way as its only one moving part.

From my experience what causes the most physics lag is collisions which is why CFrame methods are preferred as an alternative as they ignore collisions.

However best method is to go in and measure the data yourself with the tools provided like in @5uphi detailed video.

TBH its more difficult as you cannot run a simple os.clock benchmark test and get a flat number scripting wise.

Physics vs CFrame - Roblox Scripting Performance Benchmark

[5uphi]

CPU time lower is better

[Scottifly]

You could use a HingeConstraint instead of AngularVelocity, It’s much easier to set up.

That being said I’ve seen lag when you add too many physics based movers. CFraming Anchored items takes less CPU power AFAIK.