After seeing its announcement a few days ago I didn’t think much of it, but after looking into it a bit this looks awesome !
Box3D was released a few days ago by legendary developer Erin Catto accompanied by a blog post, this is the person behind the popular Box2D engine.
In the blog post he explains that his work stems from his collaboration with Dirk Gregorius, “Principal Software Engineer II and Physics Architect” at Valve, the person behind Half-Life: Alyx’s physics engine “Rubikon”, which Box3D is based off.
Facepunch has also revealed that they have been using Box3D for about a year now as well in s&box. Showcasing a cool demo
An interesting quote from the blog post:
On the Valve side, Rubikon continues to evolve and Dirk has developed optimizations (similar to those in Box3D) in a new engine called Ragnarok. Look for that in future Valve games.
👀
Did he just reveal Valve’s next physics engine ?
+HL3 confirmed
- cool showcase 1
- showcase 2
- showcase 3 in gmod (already??)
Tweet transcript
I’m happy to announce the release of a new open source 3D physics engine called Box3D. I’ve been working on this project for a few years now, but it represents over 20 years of experience writing physics engines for games. Read more here: [blog post link]
What is revolutionary about it?
Using it isn’t supporting human rights violations?
Cross platform determinism
This sounds like a big selling point to me.
From what I understand, many of the existing physics engines are non-deterministic, which can make networking physics interactions in certain types of games pretty difficult.
Difficult to say, but I see points against this being a silver bullet of any sort.
A competitive game like starcraft wouldn’t be responsive enough if they transferred all unit data every tick. So they use a lockstep model, where only player commands are transferred, and they are relying on a deterministic system (not just physics) playing out the same on all connected machines.
However, a guideline for software companies is generally don’t reinvent the wheel, but “the secret sauce” should be your own. I imagine for competitive online games, the code that makes the network interaction crisp is part of the secret sauce. Such games, StarCraft, MMORPGs etc, also tend to have simpler game-relevant physics. Colliders so characters don’t clip through each other, some simple raycasting and instant collision checks for area effects, usually.
The angle a pebble flies away from an explosion is usually not made game-relevant even in single player games - so it’s not because they can’t, it’s because it doesn’t mesh with the game.
It might open up faster online play for some indie games, but I don’t know if it will have an impact otherwise.
Yep, yep!
Pretty sure most game engines already have 3D physics engines… Which is why this news keeps falling flat. I don’t really get what’s being offered.
Is it free? Is it vastly more efficient? Is it being incorporated into one of the big three game engines?
This guy made Box2D before, a high performant 2D physics engine that is free and open source and has been used in a shit ton of 2D games. So ofcourse devs are excited with this 3D physics engine. Yes it’s free and open source as well. It’s deterministic across different CPUs, which Unreal’s Chaos physics engine and Unity’s PhysX physics engine are not. This means that simulations with the exact same starting conditions will end up having the same result no matter the device the sim is running on. This a huge benefit for multiplayer games. And benchmarks say it’s performant. Someone already implemented it in Unity and it’s faster than the built in PhysX engine in most cases.
It’s basically a free alternative to Havok. Havok is also a high performant deterministic physics engine, but costs tens of thousands to hundreds of thousands of dollars per game to use. Havok is what many triple A studios use in their games. Like Breath of the Wild uses Havok.
Alyx’s physics were genuinely groundbreaking. If this includes all the features they showed off in that game, it’ll be one hell of a contender in the physics engine market.
Though it would have been more impactful (ha) if they had released it six years ago alongside Alyx, back when the free physics options were all kind of crap compared to the commercial ones. Jolt has become the de facto free 3d physics solution within the past few years and it’s also a huge step up over the old offerings.
Edit: sadly it seems Box3D is based on a stripped-down version of Rubikon and doesn’t really do much beyond extend Box2D to the third dimension. I guess its selling point is that it’s probably much, much easier to integrate than Jolt, and that the API will be familiar to Box2D users.
Usually they aren’t very good, and proprietary way. This is fully open source and has some really good features like determinism and multi-threaded. It is much more precise and performant as well.
If you check the blog post you can see that the impetus for making this game engine was how buggy Unreal Engine’s “Chaos” physics engine was.
Its FOSS. It does some things better, specifically than Unreal. Box2D is quite famous and well supported. Read the blog post if you want to know more details.
I’ve said what I’ve said, and asked what I’ve asked. The answers aren’t in the link. “Maybe one day” is the answer.
Answers are in the link though. Unreal’s physics engine did not meet his expectations so he rolled his own.






