I've updated my benchmark for 3 active js Physics Engines.

[finscn]

I've updated my benchmark for 3 active js Physics Engines. This benchmark is very very simple and basic . But I think it has already could tell us something .

https://finscn.github.io/phy-benchmark/

Bad news , planck.js is the slowest one

[zOadT]

Thank you very much for reporting!

I will try to find and remove deoptimizations as soon as I find some time

[8Observer8]

Asus K53SV; 8 GB RAM, i3 2.2 GHertz (2 cores); Intel HD Graphics 3000; Nvidia Geforce GT 540M (1 GB); Windows 10

• box2d.ts : 27 FPS

• planck.js : 17 FPS

• matter.js : 43 FPS

[quinton-ashley]

matter.js isn't really comparable, even though it is faster the quality of its physics simulation is much worse.

[8Observer8]

I found another benchmark here: https://lusito.github.io/box2d.ts/benchmark/

I tested with laptop ASUS K53SV:

Browser: Chrome 104

Name	avg ms/frame	5th %ile	95th %ile	Ratio
box2d.js	7.53	8.099999904632568	13.600000143051147	1.00
box2d-web	14.19	15.5	12.599999904632568	1.89
@box2d/core	20.95	21.299999952316284	16.799999952316284	2.78
box2d-html5	27.44	28.200000047683716	26.399999856948853	3.65
@flyover/box2d	39.47	41.799999952316284	38	5.24
planck.js	56.85	61	44.799999952316284	7.55

[quinton-ashley]

Stats on my MacBook Air 2020 (M1 chip)

Browser: Chrome 104

Name	avg ms/frame	5th %ile	95th %ile	Ratio
box2d.js	5.43	6	6.099999999627471	1.00
box2d-web	5.87	5.599999999627471	6.1000000005587935	1.08
@box2d/core	6.35	6.099999999627471	7.2000000001862645	1.17
box2d-html5	7.07	7.3999999994412065	7	1.30
@flyover/box2d	10.50	11.199999999254942	10.200000000186265	1.93
planck.js	13.04	26.09999999962747	10.5	2.40

Browser: Firefox 104

Name	avg ms/frame	5th %ile	95th %ile	Ratio
box2d.js	5.15	6	5	1.00
box2d-web	7.12	7	6	1.38
@box2d/core	10.41	11	7	2.02
box2d-html5	10.73	12	9	2.08
planck.js	20.66	22	24	4.01
@flyover/box2d	23.38	31	23	4.54