Performance down in cannon.js v0.4.3 -> v0.6.2

[pmndrs]

Performance down in cannon.js v0.4.3 -> v0.6.2

[marcofugaro]

Thanks for the suggestion.

Your monkey patch makes the cross product evaluate to 0, and the tangents as well, thus disabling the compute tangents part of the Narrowphase? I don't know exactly what it's affecting.

Have you tried adjusting the contact equation and other world parameters instead? These ones:

cannon-es/examples/container.html

Lines 128 to 140 in 2824846

	// Tweak contact properties.
	// Contact stiffness - use to make softer/harder contacts
	world.defaultContactMaterial.contactEquationStiffness = 1e11
	// Stabilization time in number of timesteps
	world.defaultContactMaterial.contactEquationRelaxation = 2
	// Max solver iterations: Use more for better force propagation, but keep in mind that it's not very computationally cheap!
	world.solver.iterations = 10
	// Since we have many bodies and they don't move very much, we can use the less accurate quaternion normalization
	world.quatNormalizeFast = true
	world.quatNormalizeSkip = 8 // ...and we do not have to normalize every step.

[marcofugaro]

@cx20 Looks like by using those settings, the dominos behave like your original example.

Check out the live demo here: https://codesandbox.io/s/r3f-cannon-dominos-forked-ds0kv?file=/src/App.js

[cx20]

@marcofugaro Thank you for telling me the sample. My sample seemed to improve by changing the frictionEquationStiffness parameter.

https://codesandbox.io/s/r3f-cannon-dominos-forked-ds0kv?file=/src/App.js:722-752

frictionEquationStiffness: 1e2

[cx20]

Thank you for letting me know those parameters.
I will have to check the parameters a little more carefully, but I tried a little and didn't get the results I was expecting.