HTML5 best physics engine for mobile

[Salah]

Hi,

 

I started developing a HTML5 game with Box2d (without using editors), after I have developed 16 levels of the game (from total of 40 levels),

 

I found that the performance is slow on many mobile devices, I heard that Phaser physics engines are excellent on mobile, but I am

 

a little confused, should I redo the whole game on phaser framework? or I can just use the P2.js physics engine from Phaser to take

 

advantage of the code structure that I have built? or I can refine my box2d game performance without using any other framework,

 

any advice?

 

Thank You,

Salah Jaber

[chongdashu]

Are you certain that the slow performance is due to the physics - and not something else (e.g., rendering?) 

 

The reason I'm saying that is -- Phaser also has a Box2D physics engine and I'm pretty sure it works well on mobile.

[tips4design]

P2.js is not the "physics engine from phaser". It's a separate Physics engine that happens to also be included in Phaser. Here is the link to the p2.js website: https://schteppe.github.io/p2.js/

But I don't think it will be better on mobile. Depends on how many objects you have and what collision events/detection you want.

[Salah]

Are you certain that the slow performance is due to the physics - and not something else (e.g., rendering?) 

 

The reason I'm saying that is -- Phaser also has a Box2D physics engine and I'm pretty sure it works well on mobile.

I think the problem is due to physics, I have tested it without rendering, and it gives me the same FPS on Google Chrome mobile browser..

[Salah]

P2.js is not the "physics engine from phaser". It's a separate Physics engine that happens to also be included in Phaser. Here is the link to the p2.js website: https://schteppe.github.io/p2.js/

But I don't think it will be better on mobile. Depends on how many objects you have and what collision events/detection you want.

Thank you for information, then, you think that the main problem is from the scene complexity, and not from the physics engine I have used, right?

[chongdashu]

I think the problem is due to physics, I have tested it without rendering, and it gives me the same FPS on Google Chrome mobile browser..

 

I think regardless of whether it's Google Chrome's mobile browser, or some other mobile browser, rendering could still be (and I suspect probably will be) the bottleneck. For example, if you are rendering lots of different sprites on the screen without a sprite batch, the draw calls itself will probably cause a massive slowdown.

 

One thing you could do is try to do a stress test between your current implementation and Phaser. Maybe a few hundred objects of different images spinning around (or doing whatever computation you are using in your game) and compare and see if Phaser provides a performance boost. I suspect that the performance boost will not be from the physics engine (since it does use allow Box2D too).