Godot + Jolt vs Unity: Physics

🤔What is this video about❓

📡CONTACT📬

▶️PLAYLISTS🎥

🎮MY GAMES🕹️

🛜SOCIAL MEDIA🌐

🎵MUSIC🎶
➤Song: Tokyo Music Walker - Way Home
➤Patrick Patrikios
➤Alva Majo channel theme
➤Music: Evan King - CHIPTUNE ♫ 20XX
➤Zelda II The Adventure of Link Town Theme Orches
➤Haters Gonna Hate - smol edition
(Cannot find the link, if you have it please share it with me by Discord. The link is above)

It seems that this is the end of the description. If you liked the video, consider subscribing so that YouTube notifies you when I upload new useful content :)
Have a nice day!

[lettacorporation]

My Full Godot Game Development Course is now just $9.99! 🎮🔥

✅ 5+ hours of content
✅ 30+ students
✅ 6 ratings with a 4.8/5 average ⭐️⭐️⭐️⭐️⭐️

🚀 Learn how to make games with Godot 4.3, even if you're just starting out!

🎓 Enroll now:

⏳ Limited-time offer – don’t miss it!

[maymayman]

It would help if the scenes were identical. In Godot you spawn the balls above the ramp, but in Unity you spawn them below the top of the ramp. It also looks like either the angle or the FOV of the camera is slightly different.

It would be much more useful to display the number of balls instead of the time.

[lmlimpoism]

Your Godot project has better lighting which subtracts from the physics, while your Unity project has bad lighting (literally 0 shadow fidelity) assisting physics.

[mareck]

Thats not even a comparison - different shaders - gofot hbere uses alot more postprocessing - also you seem to be running in debug mode - which slows down physics quiet alot. also which jolt version is this - integrated the plugin or the new one ? because on my tests its like 10x faster than this with more than 1000s of objects

[MakDemonik]

The comparison is kind of .... bad. We are assuming the balls are spawned at similar rates but, its just a guess. If you what counts is the collider amount.
Have an object counter in the corner, and a objects spawned per second. Also while those may be default settings, they might not be equal to another.
Sometimes even a single checkbox for example, some occlusion setting can have measurable effects on the simulation. Hard to say really

[snydicus]

The lighting in the Godot example looks way more detailed / advanced

[InfinityDragon-pvyz]

It looks like the bottom layers of the balls are frozen in Unity. The same behavior can probably be configured in Godot. Yes, it's a bit more work, but it's probably not that big of a deal if you're going to do something this large. Or, what will be more correct, probably, to make the balls do not stop moving at all ... For example, some map as a funnel.

[TheCLion]

vastly different settings, very unfair comparison

are you paid by unity??

[KamikaziFly]

Looks like your unity is only pouring balls in with depth of 3 balls back, while your Jolt is pouring balls in a depth of 10 balls deep.

[xezzee]

Can't form opinion until you make them identical for test.

[TheInfamousLegend]

this is supposed to be a physics test, right?
just a few suggestions (after watching the full video too):

- make the lighting the same (not just in post-processing, but also light direction). though given you're testing physics performance solely, probably best to not have any kind of dynamically lighting in either.
- instead of using 2 different meshes to catch the balls in each scene, make one mesh that's used in both. the Godot scene mesh has a thinner base to it.
- use an identical camera setup for both, including FOV, distance, etc.
- actually change the physics so they are the same, or are pretty much the same. it's hardly a comparison if you're using default variables that don't match up because the physics engines are different.
- make sure that the method of spawning/instantiation is identical across both engines, as well as the spawn location. also mention the method you used, that's kinda important.
- should be measuring object count, not time. time is completely irrelevant here.
- might be useful to tell us the specs of your machine.

[coinopanimator]

I've been playing with unity's physics and it has some jarring limitations that need to be worked around. It's good up to a point. Mass and Friction are two that need more extreme settings.

[kevlarkevlarovisc]

Осталось понять зачем мне нужны 500 мячиков... 🤔

[HuntsmanStrong]

Yes, of course. Higher poly balls, with shadows and more realistic physics and momentum being dropped from higher up. Totally a fair and even comparison. And to shed more light, I just ran this on godot and didn’t encounter any frames issues until well after the balls began spilling out, this became less an issue of spawning objects and more that they were falling into an infinite nothing that kept having to be rendered the further the balls fell.

[Draughammer]

I think this test is invalid because you did not have the same light in both tests. You made the Unity lighting setup very bad.

Im a Godot user but you don't have to fake it like this.

[kerduslegend]

Unity 6 reworked their entire program to be ECS based. And for godot to achieve that level of performance godot need to re write their entire codebase which probably gonna took longer than what unity did

[slxyter]

this is not a valid test/comparison. shaders in Godot's editor are better/turned up more than unitys

[BetaTester]

They aren't even the same man.
If your going to make a comparison the same or equivalent variables need to be consistent

[HiveQun]

Interesting. I find that godot physics jitter and anti-aliasing settings can really affect performance as well

[avatarcity]

This comparison is not appropriate. Godot generates objects faster. Make sure the number of objects and other adjustments are equal and the versions used. Because it should be faster than this.