Page 1 of 2

Comparison of Physics Engines with Irrlicht

Posted: Sat Sep 20, 2008 4:19 pm
by Weng
I am exploring on using a physics engine with irrlicht.
So far, the more popular and free ones are PhysX, Newton and ODE.

I have visited their websites and their screenshots look pretty impressive. As this is the first time using a physics engine, I would prefer one which
has a gentler learning curve and which has an active online community,
like Irrlicht.

So among these 3 engines, which one is better in terms of picking up and
getting online help?

:)

Posted: Sat Sep 20, 2008 4:27 pm
by rogerborg

Posted: Sat Sep 20, 2008 8:51 pm
by Nadro
I also think than Bullet is currently the best free physic engine. It's very good engine with very good license :)

Posted: Sat Sep 20, 2008 9:18 pm
by kingdutch
I'll have a question to add to this topic:

Looking more FPS and realism wise (even if this meant a steeper learning curve) would Bullet still be the choice or would others be better?

Posted: Sat Sep 20, 2008 9:22 pm
by night_hawk
People know I'm sold to PhysX when it comes to physics. Not literaly, but you get the idea. It's easy to setup, easy to use and stable. Crappy version detection (correction, no version detection), but awesome overall.

Bullet seemed a bit weird to me, but that's just my opinion.

Posted: Sun Sep 21, 2008 1:57 am
by wyrmmage
Bullet is a weird physics engine, IMO, but I still like it better than, say, ODE :D I would recommend going with Bullet, but be prepared....they have very little documentation compared to more established physics engines!
-wyrmmage

Posted: Sun Sep 21, 2008 11:28 am
by Weng
Looks like Bullet is a popular physics engine.

How about physX or newton or ODE? Do they have a steeper learning curve or a less active online community?

Posted: Sun Sep 21, 2008 12:23 pm
by xDan
An annoyance I've had with ODE is that it requires a fixed timestep, if you want variable (which I assume most games would), you have to do it yourself along with some accumulator and interpolation malarkey. I don't know if other physics engines do this for you.(anyone know?)

Posted: Sun Sep 21, 2008 2:40 pm
by Dorth
physX has me, as far as I'm concerned ^^

Posted: Sun Sep 21, 2008 5:37 pm
by Dark_Kilauea
PhysX handles variable timesteps beautifully, even doing several substeps if the timestep it's given is too long.

I like PhysX because it is fairly easy to use, well documented, and very powerful.

Posted: Sun Sep 21, 2008 6:18 pm
by JP
my vote goes for Physx as it's the only one i've used extensively and it worked a dream for me and was really easy to pick up. Having said that someone else has been having a lot of trouble getting it to work at all either in Irrlicht or OpenGL code... I guess the PS3 version is just easy to use heh

Havok is another good option i'd say, commercial quality but free!

Posted: Thu Oct 23, 2008 4:36 am
by aboeing
you can use the PAL physics system to select the best physics engine for your needs. check out the PAL irrlicht benchmark demo, it compares a number of popular physics engines.
http://www.adrianboeing.com/pal/benchmark.html

Posted: Thu Oct 23, 2008 7:27 am
by CuteAlien
Wow, that looks very interesting (just can't check the demo right now, because I'm not on windows).

Also, as I'm also sometimes doing image processing, is there a working link for your ImprovCV? The one on your homepages seems to be dead :(

Posted: Thu Oct 23, 2008 2:33 pm
by CuteAlien
@aboeing: Why don't you create a project announcement for that? Especially the benchmark really rocks! I have to get the full version running some time, this is very useful stuff.

Posted: Thu Oct 23, 2008 2:35 pm
by JP
Yeah it would be useful to have a single thread with more information... might tread on the project i've just started though... :evil:

:lol:

Edit: Having checked out the benchmarker it seems cool to compare all the physics engine side by side like that but the results are pretty open to interpretation as in we don't really know what the physics simulation should look like... and if that's how the same inputs from PAL result in such varying behaviours in the resulting physics simulations then it seems PAL isn't really that good for switching between engines.. should the same inputs to PAL not result in very similar results from the different engines?