Motor Physics released

[Cortana]

The guys at Polygonal labs have finally released their motor physics engine!

Check it out:
http://lab.polygonal.de/2007/12/31/m...sics-released/

Now to see if I can rebuild his quad-tree demonstration...

[renderhjs]

isn´t that just 1 guy,- still impressive work on the performance. Even my ancient computer here still survives that,- wich it doesn´t always do with some stuff that is posted here on FK.

[Squize]

Typical, just in the past couple of days I've started a project with Box2D.

It'll be interesting to see how the performance compares to it, hopefully someone will crack and port over the Box2D examples to motor2 so we can see them side by side.

Poly guy does write some very fast code though, I'd be pretty surprised if it wasn't at least a bit quicker than Box2D.

Squize.

[ColbyCheeze]

The source seems to be missing some libraries. Luckily I already have his data structures library that it was missing but was anyone else able to find the other two library files:

de.polygonal.math2.utils.MinimumAreaRectangle;
de.polygonal.utils.DrawLib;

[newblack]

it seems to be notably faster than box2d- the attention he paid to optimization has really paid off in my opinion. it also has a much friendlier api than box2d.

there's a lack of joints; but this shouldn't be too hard to implement from here as the constraint solver is a box2d port.

i think this will prove to be the physics engine of choice for game development.

[Sietjp]

I was about to play with it. I'm petty sure this will be more performant that box2D as it has been the main focus of Michael Baczynski, and if you read the articles on his blog, you understand quickly that Mickael knows a wide range of trick to improve preformances.

[Al Capwn]

Meh. It isn't bad, but he programmed it like theoretical physics, so there is no friction.

[Sietjp]

I have the same pb as Colbycheeze, I posted a question on his blog...

[edit] You just have to comment the lines importing those classes as they are not used. It work great now

[Ray Beez]

Impressive, but too slow for game use.

[tonypa]

I doubt many games will need 1000 boxes/polygons.

[newblack]

Impressive, but too slow for game use.

was this meant sarcastically? (if so, ignore the following)- for that sort of physical simulation, i don't see anything being faster with the current player. that doesn't mean that it's the right fit for every game, or even every game with emphasis on realistic(ish) physics.

i definitely think the demos are pulling off more than the average game would require...

[Sietjp]

The sources are now available on google code but I can't make work the example with the new SVN sources.
Many modifications are done on the motor right now and I think it's better to wait for a more stable version before doing something with this negine.